In [38]:
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import scipy.stats as stats
from tqdm import tqdm
from sklearn.datasets import make_classification
from sklearn.ensemble import ExtraTreesClassifier

plt.style.use('fivethirtyeight')
sns.set(font_scale=1.5)
%config InlineBackend.figure_format = 'retina'
%matplotlib inline

1. Load data

In [4]:
conflicts=pd.read_csv('/Users/alexandra/Documents/GA_DSI/DSI5-lessons/projects/project-capstone/Armed_conflict_files/african_conflicts.csv', encoding="latin1",low_memory=False)
In [5]:
conflicts.head(3)
Out[5]:
ACTOR1 ACTOR1_ID ACTOR2 ACTOR2_ID ACTOR_DYAD_ID ADMIN1 ADMIN2 ADMIN3 ALLY_ACTOR_1 ALLY_ACTOR_2 ... INTER1 INTER2 INTERACTION LATITUDE LOCATION LONGITUDE NOTES SOURCE TIME_PRECISION YEAR
0 Police Forces of Algeria (1999-) NaN Civilians (Algeria) NaN NaN Tizi Ouzou Beni-Douala NaN NaN Berber Ethnic Group (Algeria) ... 1 7 17 36.61954 Beni Douala 4.08282 A Berber student was shot while in police cust... Associated Press Online 1 2001
1 Rioters (Algeria) NaN Police Forces of Algeria (1999-) NaN NaN Tizi Ouzou Tizi Ouzou NaN Berber Ethnic Group (Algeria) NaN ... 5 1 15 36.71183 Tizi Ouzou 4.04591 Riots were reported in numerous villages in Ka... Kabylie report 1 2001
2 Protesters (Algeria) NaN NaN NaN NaN Bejaia Amizour NaN Students (Algeria) NaN ... 6 0 60 36.64022 Amizour 4.90131 Students protested in the Amizour area. At lea... Crisis Group 1 2001

3 rows × 28 columns

In [6]:
conflicts.shape
Out[6]:
(165808, 28)
In [7]:
list(conflicts)
Out[7]:
['ACTOR1',
 'ACTOR1_ID',
 'ACTOR2',
 'ACTOR2_ID',
 'ACTOR_DYAD_ID',
 'ADMIN1',
 'ADMIN2',
 'ADMIN3',
 'ALLY_ACTOR_1',
 'ALLY_ACTOR_2',
 'COUNTRY',
 'EVENT_DATE',
 'EVENT_ID_CNTY',
 'EVENT_ID_NO_CNTY',
 'EVENT_TYPE',
 'FATALITIES',
 'GEO_PRECISION',
 'GWNO',
 'INTER1',
 'INTER2',
 'INTERACTION',
 'LATITUDE',
 'LOCATION',
 'LONGITUDE',
 'NOTES',
 'SOURCE',
 'TIME_PRECISION',
 'YEAR']
In [8]:
conflicts.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 165808 entries, 0 to 165807
Data columns (total 28 columns):
ACTOR1              165808 non-null object
ACTOR1_ID           140747 non-null float64
ACTOR2              122255 non-null object
ACTOR2_ID           140747 non-null float64
ACTOR_DYAD_ID       140747 non-null object
ADMIN1              165808 non-null object
ADMIN2              165676 non-null object
ADMIN3              86933 non-null object
ALLY_ACTOR_1        28144 non-null object
ALLY_ACTOR_2        19651 non-null object
COUNTRY             165808 non-null object
EVENT_DATE          165808 non-null object
EVENT_ID_CNTY       165808 non-null object
EVENT_ID_NO_CNTY    140747 non-null float64
EVENT_TYPE          165808 non-null object
FATALITIES          165808 non-null int64
GEO_PRECISION       165808 non-null int64
GWNO                165808 non-null int64
INTER1              165808 non-null int64
INTER2              165808 non-null int64
INTERACTION         165808 non-null int64
LATITUDE            165808 non-null object
LOCATION            165805 non-null object
LONGITUDE           165808 non-null object
NOTES               155581 non-null object
SOURCE              165635 non-null object
TIME_PRECISION      165808 non-null int64
YEAR                165808 non-null int64
dtypes: float64(3), int64(8), object(17)
memory usage: 35.4+ MB
In [9]:
conflicts.describe()
Out[9]:
ACTOR1_ID ACTOR2_ID EVENT_ID_NO_CNTY FATALITIES GEO_PRECISION GWNO INTER1 INTER2 INTERACTION TIME_PRECISION YEAR
count 140747.000000 140747.000000 140747.000000 165808.000000 165808.000000 165808.000000 165808.000000 165808.000000 165808.000000 165808.000000 165808.000000
mean 2587.796692 1211.638820 70374.000000 4.420257 1.261061 538.296964 3.385500 3.193603 32.001604 1.138751 2010.760898
std 1030.065824 1242.978894 40630.303506 76.071441 0.524248 61.974200 2.033114 2.920635 18.284381 0.441038 5.897476
min 1.000000 0.000000 1.000000 0.000000 1.000000 404.000000 1.000000 0.000000 10.000000 1.000000 1997.000000
25% 2081.000000 0.000000 35187.500000 0.000000 1.000000 490.000000 1.000000 0.000000 15.000000 1.000000 2007.000000
50% 3043.000000 675.000000 70374.000000 0.000000 1.000000 520.000000 3.000000 2.000000 28.000000 1.000000 2013.000000
75% 3288.000000 2229.000000 105560.500000 1.000000 1.000000 615.000000 5.000000 7.000000 47.000000 1.000000 2016.000000
max 3960.000000 3960.000000 140747.000000 25000.000000 3.000000 651.000000 8.000000 8.000000 88.000000 3.000000 2017.000000
In [10]:
from pandas_summary import DataFrameSummary as DFS
dfs = DFS(conflicts)
dfs.summary().T
Out[10]:
count mean std min 25% 50% 75% max counts uniques missing missing_perc types
ACTOR1 NaN NaN NaN NaN NaN NaN NaN NaN 165808 3288 0 0% categorical
ACTOR1_ID 140747 2587.8 1030.07 1 2081 3043 3288 3960 140747 3032 25061 15.11% numeric
ACTOR2 NaN NaN NaN NaN NaN NaN NaN NaN 122255 2827 43553 26.27% categorical
ACTOR2_ID 140747 1211.64 1242.98 0 0 675 2229 3960 140747 2558 25061 15.11% numeric
ACTOR_DYAD_ID NaN NaN NaN NaN NaN NaN NaN NaN 140747 9520 25061 15.11% categorical
ADMIN1 NaN NaN NaN NaN NaN NaN NaN NaN 165808 735 0 0% categorical
ADMIN2 NaN NaN NaN NaN NaN NaN NaN NaN 165676 4085 132 0.08% categorical
ADMIN3 NaN NaN NaN NaN NaN NaN NaN NaN 86933 4795 78875 47.57% categorical
ALLY_ACTOR_1 NaN NaN NaN NaN NaN NaN NaN NaN 28144 3272 137664 83.03% categorical
ALLY_ACTOR_2 NaN NaN NaN NaN NaN NaN NaN NaN 19651 2759 146157 88.15% categorical
COUNTRY NaN NaN NaN NaN NaN NaN NaN NaN 165808 50 0 0% categorical
EVENT_DATE NaN NaN NaN NaN NaN NaN NaN NaN 165808 7639 0 0% categorical
EVENT_ID_CNTY NaN NaN NaN NaN NaN NaN NaN NaN 165808 155591 0 0% categorical
EVENT_ID_NO_CNTY 140747 70374 40630.3 1 35187.5 70374 105560 140747 140747 140747 25061 15.11% numeric
EVENT_TYPE NaN NaN NaN NaN NaN NaN NaN NaN 165808 14 0 0% categorical
FATALITIES 165808 4.42026 76.0714 0 0 0 1 25000 165808 263 0 0% numeric
GEO_PRECISION 165808 1.26106 0.524248 1 1 1 1 3 165808 3 0 0% numeric
GWNO 165808 538.297 61.9742 404 490 520 615 651 165808 49 0 0% numeric
INTER1 165808 3.3855 2.03311 1 1 3 5 8 165808 8 0 0% numeric
INTER2 165808 3.1936 2.92064 0 0 2 7 8 165808 9 0 0% numeric
INTERACTION 165808 32.0016 18.2844 10 15 28 47 88 165808 43 0 0% numeric
LATITUDE NaN NaN NaN NaN NaN NaN NaN NaN 165808 19082 0 0% categorical
LOCATION NaN NaN NaN NaN NaN NaN NaN NaN 165805 17936 3 0.00% categorical
LONGITUDE NaN NaN NaN NaN NaN NaN NaN NaN 165808 19243 0 0% categorical
NOTES NaN NaN NaN NaN NaN NaN NaN NaN 155581 126318 10227 6.17% categorical
SOURCE NaN NaN NaN NaN NaN NaN NaN NaN 165635 11908 173 0.10% categorical
TIME_PRECISION 165808 1.13875 0.441038 1 1 1 1 3 165808 3 0 0% numeric
YEAR 165808 2010.76 5.89748 1997 2007 2013 2016 2017 165808 21 0 0% numeric

2. EDA - part A: Univariate analysis

A) Change to lowercase

In [11]:
conflicts.columns = [x.lower() for x in conflicts.columns]

B) Check for nulls

In [12]:
conflicts.isnull().sum()
Out[12]:
actor1                   0
actor1_id            25061
actor2               43553
actor2_id            25061
actor_dyad_id        25061
admin1                   0
admin2                 132
admin3               78875
ally_actor_1        137664
ally_actor_2        146157
country                  0
event_date               0
event_id_cnty            0
event_id_no_cnty     25061
event_type               0
fatalities               0
geo_precision            0
gwno                     0
inter1                   0
inter2                   0
interaction              0
latitude                 0
location                 3
longitude                0
notes                10227
source                 173
time_precision           0
year                     0
dtype: int64

I have some nulls but not in the columns i'm most interested in. Most likely i will not use the columns with the nulls other than hte source and the notes which i will use in the NLP section of the project. At this stage i'm not yet making a final decision about which columns to keep or remove.

Country

In [13]:
conflicts['country'].unique()
Out[13]:
array(['Algeria', 'Angola', 'Benin', 'Burkina Faso', 'Burundi',
       'Cameroon', 'Central African Republic', 'Chad',
       'Democratic Republic of Congo', 'Egypt', 'Equatorial Guinea',
       'Eritrea', 'Ethiopia', 'Gabon', 'Gambia', 'Ghana', 'Guinea',
       'Guinea-Bissau', 'Ivory Coast', 'Kenya', 'Lesotho', 'Liberia',
       'Libya', 'Madagascar', 'Malawi', 'Mali', 'Mauritania', 'Morocco',
       'Mozambique', 'Namibia', 'Niger', 'Nigeria', 'Republic of Congo',
       'Rwanda', 'Senegal', 'Sierra Leone', 'Somalia', 'South Africa',
       'South Sudan', 'Sudan', 'Swaziland', 'Tanzania', 'Togo', 'Tunisia',
       'Uganda', 'Zambia', 'Zimbabwe', 'Botswana', 'Djibouti',
       'Mozambique '], dtype=object)

**I've noticed a problem with Mozambique as it appears twice.

In [14]:
conflicts.country.replace({"Mozambique ": "Mozambique" },inplace=True)
In [15]:
len(conflicts.country.unique())
Out[15]:
49

Oh joy, so from 1996 until 2017 there have incidents of armed conflict recorded in 49 out of the 54 currently 'officially' recognized countries. Depending on who you ask africa has between 47-58 countries but i'm going with UN official number of 54. Countries with no recorded incidents include: cape verde, seychelles, comorros, sao tome. all islands and with the exception of sao tome who is too closed off, these countries have active and profitable tourism industries.

so clearly being an island with tourism potential lowers your risk of conflict. make sure you use number of neighbours as predictor as well as tourism data from world bank.

In [16]:
fig, ax = plt.subplots(figsize=(12,10))

sns.countplot(conflicts['country'], label='Count', ax=ax)
plt.xticks(rotation=90)
plt.show()

Event Type

In [17]:
conflicts.event_type.value_counts()
Out[17]:
Riots/Protests                                47224
Violence against civilians                    45021
Battle-No change of territory                 44005
Strategic development                         10923
Remote violence                               10456
Battle-Government regains territory            2632
Non-violent transfer of territory              2566
Battle-Non-state actor overtakes territory     2135
Headquarters or base established                771
Remote Violence                                  30
Violence Against Civilians                       21
Battle-no change of territory                    16
Strategic Development                             7
Strategic development                             1
Name: event_type, dtype: int64

Spelling mistakes have lead to some errors in event descriptions. Simple cleanup required.

In [18]:
conflicts.event_type.replace({"Violence Against Civilians": "Violence against civilians","Strategic Development":"Strategic development","Strategic development ":"Strategic development", "Remote Violence":"Remote violence", "Battle-no change of territory":"Battle-No change of territory"},inplace=True)
In [19]:
conflicts.event_type.value_counts()
Out[19]:
Riots/Protests                                47224
Violence against civilians                    45042
Battle-No change of territory                 44021
Strategic development                         10931
Remote violence                               10486
Battle-Government regains territory            2632
Non-violent transfer of territory              2566
Battle-Non-state actor overtakes territory     2135
Headquarters or base established                771
Name: event_type, dtype: int64
In [20]:
len(conflicts.event_type)
Out[20]:
165808
In [21]:
fig, ax = plt.subplots(figsize=(10,8))

sns.countplot(conflicts['event_type'], label='Count', ax=ax)
plt.xticks(rotation=90)
plt.show()

Actor

In [22]:
# conflicts.actor1.unique().sum()
In [23]:
conflicts.actor1.describe()
Out[23]:
count                                 165808
unique                                  3288
top       Unidentified Armed Group (Somalia)
freq                                    5896
Name: actor1, dtype: object

Clearly a large collection of different groups but they tend to fall into teh same categories so let's look at their categories. ACLED provides ua with a simplified version (inter1).

Explanation:

1- Government and State Security
2- Rebel Groups
3- Political Militias
4- Identity Militias
5- Rioters
6- Protesters
7- Civilians 
8- External/Other Forces
In [24]:
conflicts.inter1.value_counts()
Out[24]:
1    41906
3    37034
6    31610
2    21958
5    15531
4    11252
8     6477
7       40
Name: inter1, dtype: int64

Inter2 refers to whom the action was carried out on.

In [25]:
sns.countplot(conflicts['inter1'], color="b")
plt.show()
In [26]:
conflicts.inter2.value_counts()
Out[26]:
7    50765
0    43553
1    24588
2    20414
3    13655
4     6035
8     2677
5     2493
6     1628
Name: inter2, dtype: int64

Inter2 has an extra value (0) refering to the actions that didnt target a specific group

In [27]:
sns.countplot(conflicts['inter2'], color="r")
plt.show()

Interaction

In [28]:
conflicts['interaction'].describe()
Out[28]:
count    165808.000000
mean         32.001604
std          18.284381
min          10.000000
25%          15.000000
50%          28.000000
75%          47.000000
max          88.000000
Name: interaction, dtype: float64
In [29]:
conflicts['interaction'].value_counts()
Out[29]:
60    27953
37    23436
12    21299
13    14721
17    10277
27     9900
15     7279
50     6514
47     5375
16     4774
44     4331
28     4037
20     3023
30     2805
23     2420
33     2246
10     2170
38     1740
11     1706
14     1531
55     1407
18     1029
57      889
78      883
80      853
34      785
24      558
22      546
58      257
40      212
35      148
36      142
48      127
66       89
68       82
56       75
88       73
45       27
26       24
70       23
25       20
46       11
77       11
Name: interaction, dtype: int64
In [30]:
len(conflicts['interaction'].unique())
Out[30]:
43
In [31]:
fig, ax = plt.subplots(figsize=(10,8))

sns.countplot(conflicts['interaction'], label='Count', ax=ax)
plt.xticks(rotation=90)
plt.show()

Fatalities

In [32]:
conflicts.fatalities.describe()
Out[32]:
count    165808.000000
mean          4.420257
std          76.071441
min           0.000000
25%           0.000000
50%           0.000000
75%           1.000000
max       25000.000000
Name: fatalities, dtype: float64
In [33]:
conflicts.fatalities.value_counts().head(10)
Out[33]:
0     113321
1      18316
2       7978
3       4628
4       3026
5       2623
10      2541
6       1657
7       1403
8        990
Name: fatalities, dtype: int64

Fatalities need to be analysed with other features to get more meaning from this.

Year

In [32]:
fig, ax = plt.subplots(figsize=(14,8))

sns.countplot(conflicts['year'], ax=ax)
plt.show()

Bivariate analysis - Causes of conflict

In [33]:
grouped1=conflicts.groupby(['event_type', 'inter1'])
grouped1.size().head(20)
Out[33]:
event_type                                  inter1
Battle-Government regains territory         1          1875
                                            2           271
                                            3           185
                                            4             6
                                            8           295
Battle-No change of territory               1         20120
                                            2          7008
                                            3          8684
                                            4          5384
                                            8          2825
Battle-Non-state actor overtakes territory  1           803
                                            2           788
                                            3           339
                                            4            92
                                            8           113
Headquarters or base established            1            65
                                            2           318
                                            3           250
                                            4             3
                                            8           135
dtype: int64
In [34]:
event_type_actor=pd.crosstab(index=conflicts['inter1'], columns=conflicts['event_type'])
event_type_actor
Out[34]:
event_type Battle-Government regains territory Battle-No change of territory Battle-Non-state actor overtakes territory Headquarters or base established Non-violent transfer of territory Remote violence Riots/Protests Strategic development Violence against civilians
inter1
1 1875 20120 803 65 399 4187 1584 4772 8101
2 271 7008 788 318 1248 1328 3 2441 8553
3 185 8684 339 250 442 3423 48 2756 20907
4 6 5384 92 3 27 93 6 352 5289
5 0 0 0 0 0 4 14619 5 903
6 0 0 0 0 1 1 30941 13 654
7 0 0 0 0 1 2 0 37 0
8 295 2825 113 135 448 1448 23 555 635
In [35]:
event_type_actor.plot(kind='bar', figsize=(18,12), stacked=True)
Out[35]:
<matplotlib.axes._subplots.AxesSubplot at 0x10dfc1710>
In [36]:
event_type_year=pd.crosstab(index=conflicts['year'], columns=conflicts['event_type'])
event_type_year
Out[36]:
event_type Battle-Government regains territory Battle-No change of territory Battle-Non-state actor overtakes territory Headquarters or base established Non-violent transfer of territory Remote violence Riots/Protests Strategic development Violence against civilians
year
1997 152 916 159 102 219 108 403 328 958
1998 206 1286 177 133 636 120 439 371 1274
1999 208 2332 158 64 308 174 421 356 987
2000 85 1688 73 84 202 191 466 177 1278
2001 93 1322 96 27 97 110 489 216 1218
2002 75 1491 198 41 32 157 544 138 1805
2003 69 1302 126 27 4 200 512 111 1519
2004 23 1020 21 2 10 168 546 131 1351
2005 41 872 8 1 8 54 738 139 1132
2006 69 950 66 8 42 96 510 120 879
2007 35 1466 24 5 20 326 750 112 1228
2008 65 1793 54 13 55 297 791 190 1916
2009 78 1128 51 13 41 194 963 271 1282
2010 37 1435 12 51 20 303 1009 298 1403
2011 85 1760 134 40 62 610 2759 577 1800
2012 101 2333 113 36 93 791 3355 1218 2597
2013 110 3745 101 46 70 855 5360 932 3560
2014 149 4412 145 30 152 1259 5403 1007 4399
2015 219 3891 150 12 127 1264 6244 1383 4423
2016 594 6695 204 28 309 2491 12145 2369 7666
2017 138 2184 65 8 59 718 3377 487 2367
In [37]:
event_type_year.plot(kind='bar', figsize=(12,8), stacked=True)
Out[37]:
<matplotlib.axes._subplots.AxesSubplot at 0x10e0456a0>
In [38]:
event_type_year.plot(kind='line', figsize=(12,8), stacked=True)
Out[38]:
<matplotlib.axes._subplots.AxesSubplot at 0x10daa8eb8>
In [39]:
sns.factorplot(x='event_type', y='fatalities', data= conflicts, kind='bar', aspect=3).set_xticklabels(rotation=90, horizontalalignment='right', fontsize=16).set_yticklabels(fontsize=16)
plt.xlabel('Event type', fontsize=16)
plt.ylabel('Mean Fatalities', fontsize=16)
plt.show()
In [40]:
new=conflicts[['country', 'fatalities']]
new=new.groupby(['country']).sum()
new=new.sort_values(by=['fatalities'], ascending=False)

new.head(10)
Out[40]:
fatalities
country
Angola 143881
Sudan 89288
Eritrea 79399
Democratic Republic of Congo 79255
Nigeria 63685
Ethiopia 51849
Somalia 43023
South Sudan 30889
Burundi 22655
Libya 19180
In [42]:
sns.factorplot(x='year', y='fatalities', data= conflicts, kind='bar', aspect=3).set_xticklabels(rotation=90, horizontalalignment='right', fontsize=16).set_yticklabels(fontsize=16)
plt.xlabel('Year', fontsize=16)
plt.ylabel('Mean Fatalities', fontsize=16)
plt.show()

lets look at the interactions between the actors in coflict by trying to map their activities using networks

In [43]:
import matplotlib.cm as cm
import networkx as nx
from sklearn.preprocessing import MinMaxScaler
In [44]:
df = conflicts[['inter1','inter2']]
df['indicator'] = 1

/Users/alexandra/anaconda3/lib/python3.6/site-packages/ipykernel_launcher.py:2: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy
In [45]:
df_1 = df.groupby(['inter1','inter2']).count().reset_index()

Weighted networks

In [46]:
G_w = nx.DiGraph()
G_w.add_nodes_from(range(8))
G_w.add_weighted_edges_from(df_1.values)
In [47]:
print(nx.info(G_w))

Name: 
Type: DiGraph
Number of nodes: 9
Number of edges: 67
Average in degree:   7.4444
Average out degree:   7.4444
In [48]:
nx.adjacency_matrix(G_w).todense()
Out[48]:
matrix([[    0,     0,     0,     0,     0,     0,     0,     0,     0],
        [ 2170,  1707, 15033,  9027,   927,   993,  1404, 10275,   370],
        [ 3023,  6266,   545,  1062,   261,     3,     9,  9899,   890],
        [ 2805,  5693,  1359,  2246,   422,    30,    70, 23434,   975],
        [  212,   604,   297,   363,  4331,     7,     4,  5374,    60],
        [ 6515,  6286,    17,   118,    20,  1407,    39,   889,   240],
        [27952,  3370,    15,    72,     7,    36,    89,     0,    69],
        [   23,     3,     1,     2,     0,     0,     0,    11,     0],
        [  853,   659,  3147,   765,    67,    17,    13,   883,    73]],
       dtype=int64)
In [49]:
nx.pagerank(G_w)
Out[49]:
{0: 0.23689138852472078,
 1: 0.12296620852777368,
 2: 0.10768894717112877,
 3: 0.08777294799876813,
 4: 0.06504105723834373,
 5: 0.045255186922665315,
 6: 0.04301951982858464,
 7: 0.24430811839174393,
 8: 0.04705662539627082}
In [50]:
positions = nx.circular_layout(G_w)
cmap_edges = cm.get_cmap('rainbow')
edge_weights = [e[2]['weight'] for e in G_w.edges(data=True)]
In [51]:
df_1['min_max_scaled'] = MinMaxScaler().fit_transform(df_1[['indicator']])
In [52]:
df_1.sort_values('min_max_scaled', ascending=False).head(10)
Out[52]:
inter1 inter2 indicator min_max_scaled
45 6 0 27952 1.000000
25 3 7 23434 0.838360
2 1 2 15033 0.537798
7 1 7 10275 0.367572
16 2 7 9899 0.354120
3 1 3 9027 0.322922
36 5 0 6515 0.233051
37 5 1 6286 0.224858
10 2 1 6266 0.224142
19 3 1 5693 0.203642
In [53]:
nx.draw_circular(G_w,with_labels=True, edge_color=edge_weights, edge_cmap=cmap_edges)
vmin = 50
vmax = 10000
sm = plt.cm.ScalarMappable(cmap=cmap_edges, norm=plt.Normalize(vmin=vmin, vmax=vmax))
sm._A = []
plt.colorbar(sm)

plt.show()

This image is quite hard to interpret because with teh exception of node 0 all the others are interacting with each other and we can tell by the arrows that they are attacking each other back and forth. So i will have to simplify this to aid interpretation.

In [54]:
nx.draw_networkx(G_w,with_labels=True)

Now only looking at the top 20 interactions purely for interpretability

In [55]:
df_1['min_max_scaled'] = MinMaxScaler().fit_transform(df_1[['indicator']])
In [56]:
df_1.sort_values('min_max_scaled', ascending=False).head(10)
Out[56]:
inter1 inter2 indicator min_max_scaled
45 6 0 27952 1.000000
25 3 7 23434 0.838360
2 1 2 15033 0.537798
7 1 7 10275 0.367572
16 2 7 9899 0.354120
3 1 3 9027 0.322922
36 5 0 6515 0.233051
37 5 1 6286 0.224858
10 2 1 6266 0.224142
19 3 1 5693 0.203642
In [57]:
df_20=df_1.loc[df_1['min_max_scaled']>0.05]
In [58]:
df_20
Out[58]:
inter1 inter2 indicator min_max_scaled
0 1 0 2170 0.077600
1 1 1 1707 0.061035
2 1 2 15033 0.537798
3 1 3 9027 0.322922
6 1 6 1404 0.050195
7 1 7 10275 0.367572
9 2 0 3023 0.108118
10 2 1 6266 0.224142
16 2 7 9899 0.354120
18 3 0 2805 0.100318
19 3 1 5693 0.203642
21 3 3 2246 0.080319
25 3 7 23434 0.838360
31 4 4 4331 0.154914
34 4 7 5374 0.192229
36 5 0 6515 0.233051
37 5 1 6286 0.224858
41 5 5 1407 0.050302
45 6 0 27952 1.000000
46 6 1 3370 0.120532
60 8 2 3147 0.112554
In [59]:
G_20 = nx.DiGraph()
G_20.add_nodes_from(range(8))
G_20.add_weighted_edges_from(df_20.iloc[:,:3].values)
In [60]:
list_20=df_20.min_max_scaled.tolist()
list_20
Out[60]:
[0.07760008586454867,
 0.06103538334943294,
 0.5377982898644056,
 0.32292225680655434,
 0.05019498407928161,
 0.3675718221172767,
 0.10811777753926514,
 0.22414224893563736,
 0.35411970949161037,
 0.10031841436800114,
 0.2036420879396086,
 0.08031912990590677,
 0.8383599871203178,
 0.15491395656684911,
 0.19222925834496082,
 0.23305069586061322,
 0.22485778684125793,
 0.05030231476512469,
 1.0,
 0.1205323602017817,
 0.11255411255411256]
In [61]:
type(edge_weights)
Out[61]:
list
In [62]:
nx.draw_circular(G_20,with_labels=True, edge_color=list_20, edge_cmap=cmap_edges)
vmin = 50
vmax = 10000
sm = plt.cm.ScalarMappable(cmap=cmap_edges)
sm._A = []
plt.colorbar(sm)

plt.show()
In [63]:
density=nx.density(G_w)
print("Network density:", density)

Network density: 0.9305555555555556

A 0 would mean that there are no connections at all, and a 1 would indicate that all possible edges are present (a perfectly connected network. In my case a density value of 0.93 means this is a very well connected network.

In [64]:
from operator import itemgetter
In [65]:
degree_dict = dict(G_w.degree(G_w.nodes()))
nx.set_node_attributes(G_w, degree_dict, 'degree')

sorted_degree = sorted(degree_dict.items(), key=itemgetter(1), reverse=True)
sorted_degree
Out[65]:
[(1, 17),
 (2, 17),
 (3, 17),
 (4, 16),
 (5, 16),
 (8, 16),
 (6, 15),
 (7, 12),
 (0, 8)]

Feature Engineering

Creating new feature based on the number of neighbour each country has

In [66]:
list_a=conflicts['country'].unique().tolist()
In [67]:
len(list_a)
Out[67]:
49
In [68]:
list_b=[7, 4, 4, 6, 3, 6, 6, 6, 9, 4, 2, 3, 6, 3, 1, 3, 6, 2, 5, 5, 1, 3, 6, 0, 3, 7, 4, 2, 6, 4, 7, 4, 5, 4, 5, 2, 3, 6, 7, 7, 2, 8, 3, 2, 5, 7, 4, 3, 3]
In [69]:
neighbours=pd.DataFrame({'country': list_a, 'neighbours': list_b})
neighbours.head()
Out[69]:
country neighbours
0 Algeria 7
1 Angola 4
2 Benin 4
3 Burkina Faso 6
4 Burundi 3
In [70]:
conflicts =pd.merge(conflicts, neighbours, on='country', how='left')
print('nulls:', conflicts.neighbours.isnull().sum())

nulls: 0
In [71]:
conflicts.neighbours.describe()
Out[71]:
count    165808.000000
mean          4.940130
std           1.974245
min           0.000000
25%           3.000000
50%           5.000000
75%           6.000000
max           9.000000
Name: neighbours, dtype: float64
In [251]:
fig, ax = plt.subplots(figsize=(10,8))

sns.countplot(conflicts['neighbours'], label='Count', ax=ax)
plt.xticks(rotation=90)
plt.show()
In [73]:
features=conflicts[['country','year', 'inter1', 'inter2', 'interaction', 'neighbours', 'event_type', 'fatalities']]
In [75]:
features.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 165808 entries, 0 to 165807
Data columns (total 8 columns):
country        165808 non-null object
year           165808 non-null int64
inter1         165808 non-null int64
inter2         165808 non-null int64
interaction    165808 non-null int64
neighbours     165808 non-null int64
event_type     165808 non-null object
fatalities     165808 non-null int64
dtypes: int64(6), object(2)
memory usage: 16.4+ MB
In [76]:
numerical=features[['neighbours', 'year', 'fatalities']]
type(numerical)
Out[76]:
pandas.core.frame.DataFrame
In [77]:
categorical=features[['inter1', 'inter2', 'interaction', 'event_type', 'country']]
In [78]:
categorical.shape
Out[78]:
(165808, 5)
In [79]:
categorical=pd.get_dummies(categorical, columns=['inter1', 'inter2', 'interaction', 'event_type', 'country'], drop_first=True)
In [80]:
categorical.shape
Out[80]:
(165808, 113)
In [81]:
categorical.head()
Out[81]:
inter1_2 inter1_3 inter1_4 inter1_5 inter1_6 inter1_7 inter1_8 inter2_1 inter2_2 inter2_3 ... country_South Africa country_South Sudan country_Sudan country_Swaziland country_Tanzania country_Togo country_Tunisia country_Uganda country_Zambia country_Zimbabwe
0 0 0 0 0 0 0 0 0 0 0 ... 0 0 0 0 0 0 0 0 0 0
1 0 0 0 1 0 0 0 1 0 0 ... 0 0 0 0 0 0 0 0 0 0
2 0 0 0 0 1 0 0 0 0 0 ... 0 0 0 0 0 0 0 0 0 0
3 0 0 0 1 0 0 0 1 0 0 ... 0 0 0 0 0 0 0 0 0 0
4 0 0 0 1 0 0 0 1 0 0 ... 0 0 0 0 0 0 0 0 0 0

5 rows × 113 columns

Okay we've added 108 columns to the dataframe, okay thats not too bad.

In [82]:
features=pd.concat([numerical, categorical], axis=1)
features.head()
Out[82]:
neighbours year fatalities inter1_2 inter1_3 inter1_4 inter1_5 inter1_6 inter1_7 inter1_8 ... country_South Africa country_South Sudan country_Sudan country_Swaziland country_Tanzania country_Togo country_Tunisia country_Uganda country_Zambia country_Zimbabwe
0 7 2001 1 0 0 0 0 0 0 0 ... 0 0 0 0 0 0 0 0 0 0
1 7 2001 0 0 0 0 1 0 0 0 ... 0 0 0 0 0 0 0 0 0 0
2 7 2001 0 0 0 0 0 1 0 0 ... 0 0 0 0 0 0 0 0 0 0
3 7 2001 0 0 0 0 1 0 0 0 ... 0 0 0 0 0 0 0 0 0 0
4 7 2001 0 0 0 0 1 0 0 0 ... 0 0 0 0 0 0 0 0 0 0

5 rows × 116 columns

In [83]:
import sys
In [84]:
print(sys.getsizeof(features))

29285064
In [85]:
features_small=features.sample(frac=0.2, random_state=1) #random state for reproducibility
In [86]:
print(sys.getsizeof(features_small))

4808514
In [86]:
correlations_f=features.corr(method='pearson').style.format("{:.2}").background_gradient(cmap=plt.get_cmap('coolwarm'), axis=1)
correlations_f
Out[86]:
neighbours year fatalities inter1_2 inter1_3 inter1_4 inter1_5 inter1_6 inter1_7 inter1_8 inter2_1 inter2_2 inter2_3 inter2_4 inter2_5 inter2_6 inter2_7 inter2_8 interaction_11 interaction_12 interaction_13 interaction_14 interaction_15 interaction_16 interaction_17 interaction_18 interaction_20 interaction_22 interaction_23 interaction_24 interaction_25 interaction_26 interaction_27 interaction_28 interaction_30 interaction_33 interaction_34 interaction_35 interaction_36 interaction_37 interaction_38 interaction_40 interaction_44 interaction_45 interaction_46 interaction_47 interaction_48 interaction_50 interaction_55 interaction_56 interaction_57 interaction_58 interaction_60 interaction_66 interaction_68 interaction_70 interaction_77 interaction_78 interaction_80 interaction_88 event_type_Battle-No change of territory event_type_Battle-Non-state actor overtakes territory event_type_Headquarters or base established event_type_Non-violent transfer of territory event_type_Remote violence event_type_Riots/Protests event_type_Strategic development event_type_Violence against civilians country_Angola country_Benin country_Botswana country_Burkina Faso country_Burundi country_Cameroon country_Central African Republic country_Chad country_Democratic Republic of Congo country_Djibouti country_Egypt country_Equatorial Guinea country_Eritrea country_Ethiopia country_Gabon country_Gambia country_Ghana country_Guinea country_Guinea-Bissau country_Ivory Coast country_Kenya country_Lesotho country_Liberia country_Libya country_Madagascar country_Malawi country_Mali country_Mauritania country_Morocco country_Mozambique country_Namibia country_Niger country_Nigeria country_Republic of Congo country_Rwanda country_Senegal country_Sierra Leone country_Somalia country_South Africa country_South Sudan country_Sudan country_Swaziland country_Tanzania country_Togo country_Tunisia country_Uganda country_Zambia country_Zimbabwe
neighbours 1.0 0.072 0.0082 0.052 -0.018 0.011 0.017 -0.033 -0.00071 -0.076 0.039 -0.033 0.0061 0.0053 -0.0056 -0.0035 0.0046 0.023 -0.0088 0.053 -0.0065 0.0015 0.0072 -0.00083 0.014 0.011 -0.024 0.031 -0.014 -0.0058 0.0037 0.0032 0.012 -0.078 -0.011 0.011 0.023 0.0027 -0.0027 -0.013 -0.0046 0.011 0.0028 -0.002 -0.00013 0.0032 0.018 0.025 -0.015 -0.0068 0.005 0.011 -0.033 -0.0083 -0.0022 0.0027 -0.0031 -0.0074 -0.021 0.023 0.0072 0.054 -0.032 -0.082 -0.0061 -0.019 -0.022 0.029 -0.067 -0.016 -0.018 0.042 -0.21 0.04 0.084 0.036 0.58 -0.026 -0.12 -0.025 -0.049 0.095 -0.035 -0.068 -0.063 0.04 -0.056 0.0032 0.0058 -0.053 -0.09 0.11 -0.2 -0.05 0.11 -0.025 -0.14 0.039 -0.032 0.066 -0.13 0.0016 -0.029 0.0025 -0.25 -0.42 0.14 0.19 0.29 -0.051 0.1 -0.039 -0.25 0.0054 0.092 -0.092
year 0.072 1.0 -0.06 -0.26 -0.026 0.032 0.093 0.18 0.0049 -0.038 -0.041 -0.11 0.072 0.019 0.033 0.043 -0.075 -0.034 -0.0048 -0.16 0.063 0.0062 0.05 0.065 0.0064 -0.067 -0.15 -0.046 -0.055 0.012 -0.0046 -0.0037 -0.17 -0.0051 0.0037 -0.0039 0.013 -0.0018 0.011 -0.01 0.0022 0.0075 0.0035 0.0038 0.0025 0.046 -0.0041 0.054 0.041 0.0064 0.03 0.019 0.17 0.0054 0.0027 0.0077 -0.0057 -0.032 -0.045 -0.022 -0.11 -0.078 -0.079 -0.13 0.06 0.21 0.02 -0.064 -0.25 0.01 -0.0056 0.029 -0.02 0.036 0.044 -0.03 -0.042 -0.00087 0.11 -0.0072 -0.076 0.039 0.015 -0.0031 0.027 -0.042 -0.037 -0.042 -0.06 -0.022 -0.068 0.12 0.013 0.0089 0.051 0.011 0.022 0.016 -0.01 -0.0083 0.068 -0.051 -0.067 -0.02 -0.34 0.11 0.089 0.13 0.016 -0.0088 -0.0075 -0.0042 0.1 -0.15 -0.005 -0.11
fatalities 0.0082 -0.06 1.0 0.017 -0.016 0.0069 -0.016 -0.028 -0.0009 0.031 0.011 0.034 -0.0057 0.0053 -0.0048 -0.0045 -0.00051 0.014 -0.0022 0.04 -0.0079 0.00079 -0.01 -0.0085 -0.0041 0.094 -0.0066 0.0028 -0.00032 -0.0001 -0.00051 -0.00061 0.016 0.00096 -0.0076 -0.0021 0.00051 -0.0013 -0.0015 -0.012 -0.003 -0.0021 0.0062 -0.00055 -6.1e-06 0.0039 0.00084 -0.011 -0.0032 -0.0012 -0.0035 -0.0021 -0.026 -0.0013 -0.0012 -0.00068 -0.00047 0.011 -0.0041 0.0034 0.042 0.0097 -0.0039 -0.006 -0.0019 -0.035 -0.015 0.0014 0.075 -0.0018 -0.001 -0.0041 -0.0039 0.00028 -0.003 0.0042 0.0075 -0.0012 -0.0096 -0.00081 0.12 0.014 -0.002 -0.0017 -0.0031 -0.0009 -0.00015 -0.0029 -0.0068 -0.00099 -0.0042 -0.004 -0.0029 -0.0028 -0.0028 -0.0029 -0.0051 -0.0036 -0.0036 -0.0004 0.0029 -0.0021 0.0062 -0.0035 -0.0099 -0.015 -0.015 0.0037 0.011 -0.002 -0.0029 -0.002 -0.0092 -0.0034 -0.0049 -0.011
inter1_2 0.052 -0.26 0.017 1.0 -0.21 -0.11 -0.13 -0.19 -0.0061 -0.079 0.15 -0.12 -0.048 -0.051 -0.048 -0.037 0.12 0.076 -0.04 0.18 -0.12 -0.038 -0.084 -0.067 -0.1 -0.031 0.35 0.15 0.11 0.057 0.00057 0.0086 0.64 0.041 -0.051 -0.046 -0.027 -0.012 -0.011 -0.16 -0.04 -0.014 -0.064 -0.005 -0.0032 -0.072 -0.011 -0.079 -0.036 -0.0083 -0.029 -0.015 -0.18 -0.0091 -0.0087 -0.0046 -0.0032 -0.029 -0.028 -0.0082 0.047 0.08 0.056 0.13 -0.0044 -0.25 0.071 0.1 0.0066 -0.013 -0.006 -0.025 0.056 0.061 0.0071 0.0065 0.11 -0.0054 -0.081 -0.0054 -0.014 -0.024 -0.013 -0.011 -0.024 0.0033 -0.0075 -0.031 -0.049 -0.0098 0.049 -0.027 -0.031 -0.02 0.061 -0.014 -0.033 -0.029 -0.0097 0.035 -0.069 -0.02 0.0084 0.012 0.18 0.0093 -0.1 -0.0027 0.032 -0.014 -0.02 -0.015 -0.06 0.12 -0.033 -0.076
inter1_3 -0.018 -0.026 -0.016 -0.21 1.0 -0.14 -0.17 -0.26 -0.0083 -0.11 0.0082 -0.14 -0.042 -0.072 -0.063 -0.043 0.38 0.043 -0.055 -0.21 0.12 -0.052 -0.11 -0.092 -0.14 -0.042 -0.073 -0.031 0.099 -0.031 -0.0059 -0.0065 -0.14 -0.085 0.24 0.22 0.052 -0.0015 0.019 0.76 0.083 -0.019 -0.088 -0.0068 -0.0044 -0.098 -0.015 -0.11 -0.05 -0.011 -0.039 -0.021 -0.24 -0.012 -0.012 -0.0063 -0.0044 -0.039 -0.039 -0.011 -0.038 -0.018 0.017 -0.015 0.064 -0.34 0.018 0.35 -0.063 -0.012 -0.0047 -0.029 0.058 -0.014 0.085 0.011 0.014 -0.011 -0.049 -0.0037 -0.019 -0.079 -0.016 -0.0032 -0.014 -0.018 -0.013 -0.011 -0.016 0.0096 -0.037 0.029 0.012 -0.01 0.0086 -0.026 -0.041 0.054 -0.028 -0.022 0.033 0.052 0.0091 -0.024 0.01 0.057 -0.096 -0.014 0.051 -0.0064 0.0094 -0.015 -0.065 -0.067 -0.012 0.17
inter1_4 0.011 0.032 0.0069 -0.11 -0.14 1.0 -0.087 -0.13 -0.0042 -0.054 -0.072 -0.079 -0.049 0.5 -0.032 -0.026 0.1 -0.023 -0.028 -0.1 -0.084 0.13 -0.058 -0.046 -0.069 -0.021 -0.037 -0.016 -0.033 0.11 -0.003 -0.0032 -0.068 -0.043 -0.035 -0.032 0.11 -0.0081 -0.0079 -0.11 -0.028 0.13 0.61 0.0097 0.0096 0.68 0.045 -0.055 -0.025 -0.0057 -0.02 -0.011 -0.12 -0.0063 -0.006 -0.0032 -0.0022 -0.02 -0.019 -0.0057 0.13 -0.011 -0.017 -0.029 -0.061 -0.17 -0.038 0.12 -0.037 -0.00049 -0.0048 -0.0073 -0.051 -0.0066 0.0021 -0.0082 -0.027 -0.007 -0.055 -0.0045 -0.013 -0.015 -0.0091 -0.0084 0.03 -0.01 -0.0095 -0.0043 0.12 -0.0054 -0.011 0.02 -0.0022 0.0092 -0.011 -0.014 -0.024 -0.019 -0.014 -0.011 0.1 -0.014 -0.016 -0.017 -0.041 0.045 -0.02 0.053 0.025 -0.0086 0.013 -0.0082 -0.041 0.026 -0.016 -0.049
inter1_5 0.017 0.093 -0.016 -0.13 -0.17 -0.087 1.0 -0.16 -0.005 -0.065 0.23 -0.12 -0.087 -0.06 0.2 -0.024 -0.17 -0.0018 -0.033 -0.12 -0.1 -0.031 0.57 -0.055 -0.083 -0.025 -0.044 -0.018 -0.039 -0.019 0.029 -0.0039 -0.081 -0.051 -0.042 -0.038 -0.022 0.072 -0.0094 -0.13 -0.033 -0.012 -0.053 0.028 -0.0026 -0.059 -0.0089 0.63 0.29 0.031 0.23 0.11 -0.14 -0.0075 -0.0072 -0.0038 -0.0026 -0.024 -0.023 -0.0067 -0.19 -0.037 -0.022 -0.04 -0.083 0.47 -0.085 -0.15 -0.043 -0.0034 0.0093 0.038 -0.04 -0.0052 -0.0088 -0.013 -0.045 -0.0044 0.07 -0.0041 -0.016 -0.029 0.017 0.005 0.021 0.041 -0.0049 0.021 0.054 0.0015 0.0036 -0.043 0.028 0.034 -0.022 0.0063 0.01 0.018 -0.0014 0.0043 0.0029 -0.0063 -0.018 0.013 -0.047 -0.11 0.24 -0.048 -0.065 -0.0015 0.026 0.014 0.054 -0.004 0.07 -0.02
inter1_6 -0.033 0.18 -0.028 -0.19 -0.26 -0.13 -0.16 1.0 -0.0075 -0.098 -0.057 -0.18 -0.14 -0.094 -0.055 -0.034 -0.32 -0.054 -0.049 -0.19 -0.15 -0.047 -0.1 0.23 -0.12 -0.038 -0.066 -0.028 -0.059 -0.028 -0.0053 0.013 -0.12 -0.077 -0.064 -0.057 -0.033 -0.015 0.024 -0.2 -0.05 -0.017 -0.079 -0.0062 0.0092 -0.089 -0.013 -0.098 -0.045 0.016 -0.036 -0.019 0.93 0.048 0.037 -0.0057 -0.004 -0.036 -0.035 -0.01 -0.29 -0.055 -0.033 -0.061 -0.13 0.75 -0.13 -0.27 -0.046 0.043 0.015 0.068 -0.074 -0.0065 -0.044 -0.011 -0.08 -0.0008 0.094 0.0011 -0.021 0.057 0.037 0.0081 0.036 -0.0055 0.006 0.008 0.014 -0.002 0.00068 -0.032 0.029 0.016 -0.024 0.064 0.11 -0.0084 0.077 0.0015 0.036 -0.013 -0.017 0.047 -0.076 -0.15 0.2 -0.066 -0.07 0.02 0.0016 0.041 0.18 -0.046 0.036 -0.028
inter1_7 -0.00071 0.0049 -0.0009 -0.0061 -0.0083 -0.0042 -0.005 -0.0075 1.0 -0.0031 -0.0032 -0.0046 -0.0018 -0.003 -0.0019 -0.0015 -0.0011 -0.002 -0.0016 -0.006 -0.0048 -0.0015 -0.0033 -0.0027 0.00084 -0.0012 -0.0021 -0.00089 -0.0019 -0.0009 -0.00017 -0.00019 -0.0023 -0.0025 -0.002 -0.0018 -0.0011 -0.00046 -0.00045 -0.0041 -0.0016 -0.00056 -0.0025 -0.0002 -0.00013 -0.0028 -0.00043 -0.0031 -0.0014 -0.00033 -0.0011 -0.00061 -0.007 -0.00036 -0.00035 0.76 0.52 -0.0011 -0.0011 -0.00033 -0.0093 -0.0018 -0.0011 0.0012 -0.00085 -0.0098 0.054 -0.0095 -0.0022 -0.00052 -0.00028 -0.0012 0.0023 -0.0012 0.00011 0.0048 -0.0044 -0.0004 -0.0021 -0.00026 -0.00077 -0.00049 -0.00056 -0.00053 -0.00099 -0.0011 -0.00058 0.0021 -0.003 -0.00041 -0.0014 0.023 -0.0012 -0.0008 -0.0016 -0.00083 -0.0014 -0.0011 -0.001 -0.00098 -0.0044 0.014 -0.00096 0.0034 0.0091 -0.0033 -0.0041 0.0017 0.0016 -0.00054 -0.001 -0.00061 0.0022 -0.0028 -0.0014 -0.003
inter1_8 -0.076 -0.038 0.031 -0.079 -0.11 -0.054 -0.065 -0.098 -0.0031 1.0 -0.026 0.22 0.026 -0.028 -0.021 -0.016 -0.074 -0.0078 -0.021 -0.077 -0.063 -0.019 -0.043 -0.035 -0.052 0.25 -0.027 -0.012 -0.025 -0.012 -0.0022 -0.0024 -0.051 0.6 -0.026 -0.024 -0.014 -0.006 -0.0059 -0.082 0.21 -0.0072 -0.033 -0.0026 -0.0016 -0.037 0.07 -0.041 -0.019 -0.0043 -0.015 0.0055 -0.091 -0.0047 0.014 -0.0024 -0.0016 0.36 0.36 0.1 0.078 0.0082 0.048 0.088 0.13 -0.13 0.016 -0.079 -0.024 -0.0068 -0.0019 -0.015 -0.039 -0.01 0.023 0.0039 0.011 -0.004 -0.047 -0.0034 0.16 -0.03 -0.0064 -0.0031 -0.012 -0.011 -0.00092 -0.0067 -0.028 0.004 -0.0029 0.0057 -0.016 -0.01 0.037 -0.0084 -0.014 -0.012 -0.0034 -0.0063 -0.051 -0.0032 -0.0034 -0.016 0.11 0.19 -0.051 -0.0054 -0.027 -0.007 -0.0098 -0.0072 -0.033 -0.027 -0.014 -0.039
inter2_1 0.039 -0.041 0.011 0.15 0.0082 -0.072 0.23 -0.057 -0.0032 -0.026 1.0 -0.16 -0.13 -0.081 -0.052 -0.042 -0.28 -0.053 0.24 0.16 0.21 0.067 0.43 0.27 -0.11 0.11 -0.057 -0.024 -0.051 -0.024 -0.0046 -0.005 -0.11 -0.066 -0.055 -0.049 -0.029 -0.012 -0.012 -0.17 -0.043 -0.015 -0.068 -0.0053 -0.0034 -0.076 -0.012 -0.084 -0.039 -0.0089 -0.031 -0.016 -0.19 -0.0097 -0.0093 -0.0049 -0.0034 -0.031 -0.03 -0.0088 0.21 0.077 -0.027 -0.046 -0.023 0.077 -0.071 -0.23 -0.047 -0.0044 0.002 0.0054 0.00018 0.031 -0.037 0.029 0.0099 -0.0011 0.029 -0.0019 0.06 0.0028 0.013 -0.0021 -0.0031 0.021 0.031 -0.0079 0.00093 0.0092 0.037 -0.031 0.026 0.0052 0.011 0.03 0.017 0.022 -0.017 0.039 -0.044 0.049 -0.0068 0.0066 -0.065 -0.015 0.012 0.0095 0.02 0.0082 0.0019 0.016 -0.0096 0.024 -0.0043 -0.049
inter2_2 -0.033 -0.11 0.034 -0.12 -0.14 -0.079 -0.12 -0.18 -0.0046 0.22 -0.16 1.0 -0.11 -0.073 -0.046 -0.037 -0.25 -0.048 -0.038 0.68 -0.12 -0.036 -0.08 -0.065 -0.096 -0.03 -0.051 0.15 0.16 0.072 0.024 0.018 -0.094 0.32 -0.049 -0.044 -0.026 -0.011 -0.011 -0.15 -0.039 -0.013 -0.061 -0.0048 -0.0031 -0.069 -0.01 -0.076 -0.035 -0.008 -0.028 -0.015 -0.17 -0.0087 -0.0083 -0.0044 -0.0031 -0.027 -0.027 -0.0079 0.4 0.11 -0.023 0.0018 0.044 -0.24 -0.059 -0.23 0.23 -0.013 -0.0067 -0.028 -0.0073 -0.017 -0.041 0.0088 0.027 0.0065 -0.067 -0.0018 -0.016 0.036 -0.013 -0.012 -0.024 -0.016 -0.0082 -0.017 -0.049 -0.01 0.017 -0.0072 -0.03 -0.019 0.0054 -0.014 -0.032 -0.027 -0.01 -0.0094 -0.033 -0.02 -0.0015 -0.0014 0.038 0.16 -0.099 0.051 -0.023 -0.013 -0.024 -0.015 -0.061 0.022 -0.033 -0.073
inter2_3 0.0061 0.072 -0.0057 -0.048 -0.042 -0.049 -0.087 -0.14 -0.0018 0.026 -0.13 -0.11 1.0 -0.058 -0.037 -0.03 -0.2 -0.038 -0.031 -0.12 0.6 -0.029 -0.064 -0.052 -0.077 -0.024 -0.041 -0.017 0.16 -0.017 -0.0033 -0.0036 -0.075 -0.047 -0.039 0.39 0.095 0.078 0.045 -0.12 0.13 -0.011 -0.049 -0.0038 -0.0024 -0.055 -0.0083 -0.061 -0.028 -0.0064 -0.022 -0.012 -0.13 -0.0069 -0.0067 -0.0035 -0.0024 -0.022 -0.022 -0.0063 0.31 0.022 -0.012 -0.031 0.088 -0.18 0.011 -0.18 -0.037 -0.0081 -0.0041 -0.016 -0.031 -0.011 0.029 -0.011 0.05 -0.0053 0.083 -0.0037 -0.013 -0.043 -0.01 -0.0069 -0.013 -0.008 -0.0083 0.015 -0.018 -0.0072 -0.021 0.13 -0.0079 -0.011 0.00035 -0.014 -0.02 0.0021 -0.016 0.0079 0.005 -0.0045 -0.0082 -0.017 -0.04 0.093 -0.073 -0.031 -0.048 -0.01 -0.008 -0.0096 -0.015 -0.045 -0.017 -0.045
inter2_4 0.0053 0.019 0.0053 -0.051 -0.072 0.5 -0.06 -0.094 -0.003 -0.028 -0.081 -0.073 -0.058 1.0 -0.024 -0.019 -0.13 -0.025 -0.02 -0.075 -0.061 0.29 -0.042 -0.033 -0.05 -0.015 -0.026 -0.011 -0.024 0.13 -0.0021 -0.0023 -0.049 -0.031 -0.025 -0.023 0.18 -0.0058 -0.0057 -0.079 -0.02 -0.007 0.84 0.048 0.026 -0.036 0.073 -0.039 -0.018 -0.0041 -0.014 -0.0077 -0.088 -0.0045 -0.0043 -0.0023 -0.0016 -0.014 -0.014 -0.0041 0.29 0.0018 -0.013 -0.021 -0.032 -0.12 -0.038 -0.12 -0.026 0.0021 -0.0035 -0.0051 -0.041 0.014 0.00071 -0.0032 -0.023 -0.0038 -0.038 -0.0032 -0.009 0.0097 -0.007 -0.0056 0.025 -0.0042 -0.0064 0.0048 0.051 -0.004 -0.01 0.063 -0.00029 -0.01 -0.004 -0.0097 -0.018 -0.014 -0.011 -0.0031 0.05 -0.01 -0.012 -0.012 -0.032 0.056 -0.048 0.036 0.023 -0.0067 0.0014 -0.0044 -0.026 0.0052 -0.017 -0.037
inter2_5 -0.0056 0.033 -0.0048 -0.048 -0.063 -0.032 0.2 -0.055 -0.0019 -0.021 -0.052 -0.046 -0.037 -0.024 1.0 -0.012 -0.082 -0.016 -0.013 -0.047 -0.039 -0.012 0.21 -0.021 -0.032 -0.0098 -0.017 -0.0071 -0.015 -0.0072 0.012 -0.0015 -0.031 -0.02 -0.016 -0.014 -0.0085 0.046 -0.0036 -0.05 -0.013 -0.0044 -0.02 0.026 -0.001 -0.023 -0.0034 -0.025 0.75 0.081 -0.0091 0.017 -0.056 -0.0029 -0.0027 -0.0015 -0.001 -0.009 -0.0089 -0.0026 -0.074 -0.014 -0.0084 -0.015 -0.032 0.19 -0.032 -0.075 -0.016 -0.0027 0.0033 0.0028 -0.021 -0.0073 -0.0061 -0.0082 -0.025 0.0025 0.086 -0.0021 -0.0051 0.0018 -0.00032 0.0075 0.017 0.037 -0.0046 0.011 0.02 -0.0014 -0.0036 -0.021 0.00062 0.024 -0.01 -0.0028 0.0053 0.0025 -0.0039 -0.0047 0.0036 -0.0057 -0.0076 0.012 -0.014 -0.047 0.031 -0.016 -0.022 3.3e-05 0.0072 0.00015 0.011 0.016 0.021 0.0039
inter2_6 -0.0035 0.043 -0.0045 -0.037 -0.043 -0.026 -0.024 -0.034 -0.0015 -0.016 -0.042 -0.037 -0.03 -0.019 -0.012 1.0 -0.066 -0.013 -0.01 -0.038 -0.031 -0.0096 -0.021 0.5 -0.026 -0.0079 -0.014 -0.0057 -0.012 -0.0058 -0.0011 0.045 -0.025 -0.016 -0.013 -0.012 -0.0069 -0.003 0.14 -0.04 -0.01 -0.0036 -0.016 -0.0013 0.029 -0.018 -0.0028 -0.02 -0.0092 0.11 -0.0073 -0.0039 -0.045 0.23 0.034 -0.0012 -0.00081 -0.0073 -0.0072 -0.0021 -0.06 -0.011 -0.0068 -0.012 -0.024 0.036 -0.021 0.058 -0.013 0.0058 -0.0018 -0.0055 -0.011 -0.0042 -0.014 -0.0057 -0.024 0.0045 0.04 0.0094 -0.0049 0.16 -0.00019 0.0039 -0.0035 0.011 -0.0021 0.011 -0.007 0.0019 -0.0017 -0.013 -0.0011 0.00088 -0.0091 -0.0018 0.047 -0.0023 -0.004 -0.0053 -0.015 -0.003 -0.0062 -0.0045 -0.016 -0.037 -0.01 -0.017 -0.0017 0.041 0.0024 0.016 0.014 0.0046 -0.0025 0.001
inter2_7 0.0046 -0.075 -0.00051 0.12 0.38 0.1 -0.17 -0.32 -0.0011 -0.074 -0.28 -0.25 -0.2 -0.13 -0.082 -0.066 1.0 -0.085 -0.068 -0.26 -0.21 -0.064 -0.14 -0.11 0.39 -0.052 -0.091 -0.038 -0.081 -0.039 -0.0073 -0.008 0.38 -0.1 -0.087 -0.078 -0.046 -0.02 -0.019 0.61 -0.068 -0.024 -0.11 -0.0085 -0.0054 0.28 -0.018 -0.13 -0.061 -0.014 0.11 -0.026 -0.3 -0.015 -0.015 -0.0078 0.012 0.11 -0.048 -0.014 -0.4 -0.076 -0.045 -0.081 0.026 -0.42 0.01 0.87 -0.04 -0.014 -0.0075 -0.033 0.12 0.0035 0.055 0.0027 0.0098 0.0059 -0.093 0.0062 -0.00077 -0.056 -0.016 0.0073 -0.012 -0.0044 -0.013 -0.0079 0.0091 0.0039 -0.02 -0.05 -0.01 0.0021 -0.015 -0.025 -0.043 0.015 -0.02 -0.019 0.021 -0.0061 0.037 -0.019 0.043 -0.02 -0.098 0.032 0.1 0.00014 0.013 -0.017 -0.092 0.025 -0.0074 0.17
inter2_8 0.023 -0.034 0.014 0.076 0.043 -0.023 -0.0018 -0.054 -0.002 -0.0078 -0.053 -0.048 -0.038 -0.025 -0.016 -0.013 -0.085 1.0 -0.013 -0.049 -0.04 -0.012 -0.027 -0.022 -0.033 0.22 -0.017 -0.0074 -0.016 -0.0074 -0.0014 -0.0015 -0.032 0.26 -0.017 -0.015 -0.0088 -0.0038 -0.0038 -0.052 0.44 -0.0046 -0.021 -0.0016 -0.001 -0.023 0.1 -0.026 -0.012 -0.0027 -0.0094 0.29 -0.058 -0.003 0.15 -0.0015 -0.001 -0.0094 -0.0092 0.16 0.11 -0.007 -0.0066 -0.0064 0.043 -0.05 -0.0061 -0.077 -0.0057 -0.0043 0.00039 -0.0071 -0.017 -0.0026 0.068 0.0052 0.028 -0.0015 -0.017 -0.0021 -0.0015 -0.0031 -0.0046 0.007 -0.0074 -0.0055 0.0003 0.0046 -0.02 0.0056 0.014 -0.007 -0.0085 -0.0056 0.075 -0.0032 -0.0079 -0.0068 -0.0065 -0.0066 -0.033 -0.0024 -0.0017 -0.0077 -0.017 0.06 -0.018 -0.003 0.0087 -0.00027 -0.0065 -0.0051 -0.02 -0.016 -0.0092 -0.024
interaction_11 -0.0088 -0.0048 -0.0022 -0.04 -0.055 -0.028 -0.033 -0.049 -0.0016 -0.021 0.24 -0.038 -0.031 -0.02 -0.013 -0.01 -0.068 -0.013 1.0 -0.039 -0.032 -0.0098 -0.022 -0.018 -0.026 -0.0081 -0.014 -0.0059 -0.012 -0.0059 -0.0011 -0.0012 -0.026 -0.016 -0.013 -0.012 -0.007 -0.003 -0.003 -0.041 -0.011 -0.0036 -0.017 -0.0013 -0.00083 -0.019 -0.0028 -0.021 -0.0094 -0.0022 -0.0075 -0.004 -0.046 -0.0024 -0.0023 -0.0012 -0.00083 -0.0075 -0.0073 -0.0021 0.13 0.044 -0.007 -0.011 -0.024 -0.064 0.0025 -0.062 -0.013 0.00014 -0.0018 0.0064 -0.0058 -0.0077 0.0055 -0.00048 0.007 0.0043 -0.023 0.0019 -0.0014 -0.0041 -0.0037 0.012 -0.0056 0.012 0.12 0.011 -0.016 0.022 -0.0073 -0.02 -0.0037 -0.0052 -0.0012 0.0025 -0.0093 -0.0075 -0.0069 0.051 -0.018 0.0035 0.0024 -0.0085 -0.017 0.07 -0.026 0.056 -0.0082 -0.0018 -0.0069 -0.004 -0.016 -0.012 -0.0077 -0.017
interaction_12 0.053 -0.16 0.04 0.18 -0.21 -0.1 -0.12 -0.19 -0.006 -0.077 0.16 0.68 -0.12 -0.075 -0.047 -0.038 -0.26 -0.049 -0.039 1.0 -0.12 -0.037 -0.082 -0.066 -0.099 -0.03 -0.052 -0.022 -0.047 -0.022 -0.0042 -0.0046 -0.097 -0.061 -0.05 -0.045 -0.026 -0.011 -0.011 -0.16 -0.04 -0.014 -0.063 -0.0049 -0.0031 -0.07 -0.011 -0.078 -0.036 -0.0082 -0.028 -0.015 -0.17 -0.0089 -0.0085 -0.0045 -0.0031 -0.028 -0.028 -0.0081 0.43 0.15 -0.024 -0.018 0.01 -0.24 -0.062 -0.23 0.23 -0.013 -0.0069 -0.027 0.03 0.017 -0.042 0.026 0.028 0.0081 -0.066 -0.0021 -0.015 0.069 -0.014 -0.012 -0.024 -0.013 -0.0057 -0.026 -0.048 -0.01 0.066 -0.015 -0.031 -0.02 0.001 -0.011 -0.032 -0.028 -0.015 0.016 -0.04 -0.021 0.0046 0.0047 -0.054 0.045 -0.1 0.078 0.039 -0.013 -0.023 -0.015 -0.059 0.071 -0.034 -0.074
interaction_13 -0.0065 0.063 -0.0079 -0.12 0.12 -0.084 -0.1 -0.15 -0.0048 -0.063 0.21 -0.12 0.6 -0.061 -0.039 -0.031 -0.21 -0.04 -0.032 -0.12 1.0 -0.03 -0.067 -0.054 -0.08 -0.025 -0.043 -0.018 -0.038 -0.018 -0.0034 -0.0038 -0.079 -0.049 -0.041 -0.037 -0.022 -0.0093 -0.0091 -0.13 -0.032 -0.011 -0.051 -0.004 -0.0025 -0.057 -0.0086 -0.063 -0.029 -0.0066 -0.023 -0.012 -0.14 -0.0072 -0.0069 -0.0037 -0.0025 -0.023 -0.022 -0.0066 0.33 0.0037 -0.021 -0.032 0.091 -0.2 0.016 -0.19 -0.039 -0.0086 -0.0044 -0.017 -0.016 -0.0018 -0.028 0.0077 0.047 -0.0057 0.11 -0.0014 -0.011 -0.045 -0.0095 -0.0087 -0.015 -0.0036 -0.006 0.0096 -0.018 -0.00038 -0.026 0.091 0.014 -0.014 0.0017 -0.014 -0.02 0.038 -0.02 -0.0059 0.0097 0.063 -0.0018 -0.019 -0.048 0.086 -0.076 -0.025 -0.054 -0.0059 -0.0093 -0.011 -0.0077 -0.043 -0.023 -0.057
interaction_14 0.0015 0.0062 0.00079 -0.038 -0.052 0.13 -0.031 -0.047 -0.0015 -0.019 0.067 -0.036 -0.029 0.29 -0.012 -0.0096 -0.064 -0.012 -0.0098 -0.037 -0.03 1.0 -0.021 -0.017 -0.025 -0.0076 -0.013 -0.0055 -0.012 -0.0056 -0.0011 -0.0012 -0.024 -0.015 -0.013 -0.011 -0.0067 -0.0029 -0.0028 -0.039 -0.0099 -0.0035 -0.016 -0.0012 -0.00079 -0.018 -0.0027 -0.02 -0.0089 -0.0021 -0.0071 -0.0038 -0.043 -0.0022 -0.0021 -0.0011 -0.00079 -0.0071 -0.0069 -0.002 0.14 -0.0026 -0.0066 -0.008 -0.0072 -0.061 -0.014 -0.059 -0.012 -0.0032 -0.0017 -0.0036 -0.021 -0.0048 -0.011 -0.0045 -0.018 -8.3e-05 -0.018 -0.0016 -0.0035 0.022 -0.0035 -0.0033 -0.0042 -0.0062 -0.0036 -0.0034 0.044 -0.0026 -0.0053 0.022 -0.0062 -0.0025 -0.0075 -0.0051 -0.0088 -0.0062 -0.0065 -0.0021 0.013 -0.0052 -0.006 -0.0065 -0.014 0.021 -0.022 0.026 0.00049 -0.0033 0.001 -0.0038 -0.013 0.056 -0.0085 -0.018
interaction_15 0.0072 0.05 -0.01 -0.084 -0.11 -0.058 0.57 -0.1 -0.0033 -0.043 0.43 -0.08 -0.064 -0.042 0.21 -0.021 -0.14 -0.027 -0.022 -0.082 -0.067 -0.021 1.0 -0.037 -0.055 -0.017 -0.029 -0.012 -0.026 -0.012 -0.0024 -0.0026 -0.054 -0.034 -0.028 -0.025 -0.015 -0.0064 -0.0063 -0.087 -0.022 -0.0077 -0.035 -0.0027 -0.0017 -0.039 -0.0059 -0.043 -0.02 -0.0046 -0.016 -0.0084 -0.096 -0.005 -0.0048 -0.0025 -0.0017 -0.016 -0.015 -0.0045 -0.13 -0.024 -0.015 -0.027 -0.056 0.34 -0.057 -0.13 -0.028 0.0033 0.0028 0.015 -0.024 0.0035 -0.027 -0.0058 -0.024 0.0035 0.067 -0.0018 -0.011 -0.012 0.019 0.0014 0.016 0.042 -0.0033 0.0099 0.035 -0.0013 0.0029 -0.037 0.032 0.033 -0.017 0.017 0.023 0.0093 -0.0014 0.015 -0.026 -0.0032 -0.011 0.012 -0.03 -0.075 0.14 -0.035 -0.039 0.0062 0.018 0.019 0.035 0.01 0.03 -0.016
interaction_16 -0.00083 0.065 -0.0085 -0.067 -0.092 -0.046 -0.055 0.23 -0.0027 -0.035 0.27 -0.065 -0.052 -0.033 -0.021 0.5 -0.11 -0.022 -0.018 -0.066 -0.054 -0.017 -0.037 1.0 -0.044 -0.014 -0.023 -0.0099 -0.021 -0.01 -0.0019 -0.0021 -0.043 -0.027 -0.023 -0.02 -0.012 -0.0051 -0.005 -0.07 -0.018 -0.0062 -0.028 -0.0022 -0.0014 -0.032 -0.0048 -0.035 -0.016 -0.0037 -0.013 -0.0068 -0.078 -0.004 -0.0038 -0.002 -0.0014 -0.013 -0.012 -0.0036 -0.1 -0.02 -0.012 -0.022 -0.045 0.16 -0.042 0.0065 -0.0088 0.006 0.005 0.0079 -0.018 0.0024 -0.024 0.0059 -0.022 0.0039 0.056 0.0058 -0.0086 0.12 0.022 0.0038 -0.00075 0.011 -0.00068 0.007 0.0033 -0.00054 0.00023 -0.03 0.0087 0.0088 -0.013 0.055 0.077 -0.0042 -0.0036 -0.0052 -0.025 -0.00042 -0.0089 0.0054 -0.028 -0.057 -0.0027 -0.028 0.011 0.045 0.0094 0.037 0.0075 0.0016 0.0036 0.012
interaction_17 0.014 0.0064 -0.0041 -0.1 -0.14 -0.069 -0.083 -0.12 0.00084 -0.052 -0.11 -0.096 -0.077 -0.05 -0.032 -0.026 0.39 -0.033 -0.026 -0.099 -0.08 -0.025 -0.055 -0.044 1.0 -0.02 -0.035 -0.015 -0.031 -0.015 -0.0028 -0.0031 -0.065 -0.041 -0.034 -0.03 -0.018 -0.0077 -0.0075 -0.1 -0.026 -0.0092 -0.042 -0.0033 -0.0021 -0.047 -0.0071 -0.052 -0.024 -0.0055 -0.019 -0.01 -0.12 -0.006 -0.0057 -0.003 -0.0021 -0.019 -0.018 -0.0054 -0.15 -0.029 -0.018 -0.032 0.051 -0.16 0.11 0.25 -0.0084 -0.0049 -0.0004 -0.0083 0.11 -0.004 -0.018 0.0036 -0.021 0.029 -0.0089 0.021 0.021 0.038 0.0026 0.02 -0.0077 0.01 0.001 0.024 -0.026 0.0025 0.0023 -0.033 -0.0031 0.0035 -0.011 -0.0019 -0.0067 0.0021 -0.0081 -0.0099 -0.04 0.0027 0.04 -0.012 -0.035 -0.059 -0.058 0.057 0.093 0.018 0.0059 0.00064 -0.032 -0.009 0.0032 0.11
interaction_18 0.011 -0.067 0.094 -0.031 -0.042 -0.021 -0.025 -0.038 -0.0012 0.25 0.11 -0.03 -0.024 -0.015 -0.0098 -0.0079 -0.052 0.22 -0.0081 -0.03 -0.025 -0.0076 -0.017 -0.014 -0.02 1.0 -0.011 -0.0045 -0.0096 -0.0046 -0.00087 -0.00095 -0.02 -0.012 -0.01 -0.0093 -0.0055 -0.0024 -0.0023 -0.032 -0.0081 -0.0028 -0.013 -0.001 -0.00064 -0.014 -0.0022 -0.016 -0.0073 -0.0017 -0.0058 -0.0031 -0.036 -0.0018 -0.0018 -0.00093 -0.00064 -0.0058 -0.0057 -0.0017 0.075 0.0012 -0.0031 0.0032 0.043 -0.05 0.0022 -0.048 0.0097 -0.0027 0.0029 -0.0043 -0.012 0.0074 0.0042 0.026 -0.0039 0.0009 -0.018 -0.0013 0.34 0.012 -0.0028 0.0019 -0.005 -0.0048 0.0052 0.0043 -0.0092 0.027 0.012 0.06 -0.0064 -0.004 -0.0082 -0.0013 -0.0063 -0.0037 -0.0042 -0.0013 -0.021 0.0015 0.0076 -0.0056 -0.013 -0.024 -0.02 0.017 0.013 -0.0027 -0.003 -0.0012 -0.012 -0.0058 -0.0061 -0.015
interaction_20 -0.024 -0.15 -0.0066 0.35 -0.073 -0.037 -0.044 -0.066 -0.0021 -0.027 -0.057 -0.051 -0.041 -0.026 -0.017 -0.014 -0.091 -0.017 -0.014 -0.052 -0.043 -0.013 -0.029 -0.023 -0.035 -0.011 1.0 -0.0078 -0.017 -0.0079 -0.0015 -0.0016 -0.034 -0.022 -0.018 -0.016 -0.0094 -0.0041 -0.004 -0.055 -0.014 -0.0049 -0.022 -0.0017 -0.0011 -0.025 -0.0038 -0.028 -0.013 -0.0029 -0.01 -0.0054 -0.061 -0.0032 -0.003 -0.0016 -0.0011 -0.01 -0.0098 -0.0029 -0.082 -0.016 0.2 0.42 -0.017 -0.086 0.22 -0.083 0.031 -0.0046 -0.0024 -0.0085 -0.0098 0.0018 -0.0057 -0.0043 0.046 -0.0035 -0.029 -0.0023 -0.0031 -0.017 -0.0024 -0.0033 -0.0087 0.0072 -0.0015 -0.014 -0.015 -0.0036 -0.007 -0.013 -0.011 -0.007 0.019 -0.0064 -0.0094 -0.01 -0.0045 -0.0043 -0.03 -0.0064 0.0055 -0.0053 0.22 -0.0074 -0.036 -0.0083 -0.017 -0.0047 -0.0044 -0.0054 -0.022 0.057 -0.012 -0.026
interaction_22 0.031 -0.046 0.0028 0.15 -0.031 -0.016 -0.018 -0.028 -0.00089 -0.012 -0.024 0.15 -0.017 -0.011 -0.0071 -0.0057 -0.038 -0.0074 -0.0059 -0.022 -0.018 -0.0055 -0.012 -0.0099 -0.015 -0.0045 -0.0078 1.0 -0.007 -0.0033 -0.00063 -0.00069 -0.014 -0.0091 -0.0075 -0.0067 -0.004 -0.0017 -0.0017 -0.023 -0.0059 -0.0021 -0.0094 -0.00073 -0.00047 -0.011 -0.0016 -0.012 -0.0053 -0.0012 -0.0042 -0.0023 -0.026 -0.0013 -0.0013 -0.00068 -0.00047 -0.0042 -0.0041 -0.0012 0.071 0.056 -0.0024 -0.0064 -0.014 -0.036 -0.0025 -0.035 -0.008 -0.0019 -0.001 -0.0045 0.009 -0.0043 0.00063 0.0025 0.049 -0.0015 -0.014 -0.00096 -0.0029 -0.01 -0.0021 -0.0019 -0.0037 -0.0042 0.00065 0.0092 -0.0069 -0.0015 0.0052 -0.012 -0.0046 -0.0029 0.032 -0.0031 -0.0052 -0.0042 0.016 -0.0036 -0.013 -0.0031 -0.0036 0.043 0.0042 0.0011 -0.015 0.00064 0.011 -0.002 -0.0039 -0.0023 -0.0096 -0.0097 -0.0051 -0.011
interaction_23 -0.014 -0.055 -0.00032 0.11 0.099 -0.033 -0.039 -0.059 -0.0019 -0.025 -0.051 0.16 0.16 -0.024 -0.015 -0.012 -0.081 -0.016 -0.012 -0.047 -0.038 -0.012 -0.026 -0.021 -0.031 -0.0096 -0.017 -0.007 1.0 -0.0071 -0.0013 -0.0015 -0.031 -0.019 -0.016 -0.014 -0.0084 -0.0036 -0.0036 -0.049 -0.013 -0.0044 -0.02 -0.0016 -0.00099 -0.022 -0.0034 -0.025 -0.011 -0.0026 -0.0089 -0.0048 -0.055 -0.0028 -0.0027 -0.0014 -0.00099 -0.0089 -0.0088 -0.0026 0.13 0.097 -0.0061 -0.011 0.012 -0.077 -0.021 -0.074 -0.017 -0.0041 -0.0022 -0.0077 -0.017 -0.0092 -0.011 -0.0051 0.033 -0.0032 -0.029 -0.002 -0.006 -0.019 -0.0044 -0.0041 -0.0078 -0.0055 -0.0046 -0.0026 -0.022 -0.0032 -0.0061 0.042 -0.0098 -0.0062 0.023 -0.0065 -0.01 -0.0089 -0.0082 0.029 -0.026 -0.0065 -0.0075 -0.01 0.1 0.073 -0.032 -0.007 -0.0034 -0.0042 -0.0082 -0.0048 -0.019 -0.02 -0.011 -0.024
interaction_24 -0.0058 0.012 -0.0001 0.057 -0.031 0.11 -0.019 -0.028 -0.0009 -0.012 -0.024 0.072 -0.017 0.13 -0.0072 -0.0058 -0.039 -0.0074 -0.0059 -0.022 -0.018 -0.0056 -0.012 -0.01 -0.015 -0.0046 -0.0079 -0.0033 -0.0071 1.0 -0.00064 -0.0007 -0.015 -0.0092 -0.0076 -0.0068 -0.004 -0.0017 -0.0017 -0.024 -0.006 -0.0021 -0.0095 -0.00074 -0.00047 -0.011 -0.0016 -0.012 -0.0054 -0.0012 -0.0043 -0.0023 -0.026 -0.0013 -0.0013 -0.00068 -0.00047 -0.0043 -0.0042 -0.0012 0.077 0.026 -0.004 -0.0022 -0.0053 -0.037 -0.01 -0.035 -0.0081 -0.002 -0.001 -0.0046 -0.012 0.05 0.0045 -0.00073 -0.0077 -0.0015 -0.011 -0.00097 -0.0029 -0.0097 -0.0021 -0.002 -0.0037 -0.0043 -0.0022 -0.0061 -0.0099 -0.0016 0.026 0.039 -0.0047 -0.003 0.0042 -0.0031 -0.0053 -0.0043 -0.0039 0.0046 -0.0045 -0.0031 -0.0036 -0.0048 0.0033 0.03 -0.015 -0.00079 0.00062 -0.002 -0.0039 -0.0023 -0.0097 0.0051 -0.0051 -0.011
interaction_25 0.0037 -0.0046 -0.00051 0.00057 -0.0059 -0.003 0.029 -0.0053 -0.00017 -0.0022 -0.0046 0.024 -0.0033 -0.0021 0.012 -0.0011 -0.0073 -0.0014 -0.0011 -0.0042 -0.0034 -0.0011 -0.0024 -0.0019 -0.0028 -0.00087 -0.0015 -0.00063 -0.0013 -0.00064 1.0 -0.00013 -0.0028 -0.0017 -0.0014 -0.0013 -0.00076 -0.00033 -0.00032 -0.0045 -0.0011 -0.00039 -0.0018 -0.00014 -8.9e-05 -0.002 -0.0003 -0.0022 -0.001 -0.00023 -0.00081 -0.00043 -0.0049 -0.00025 -0.00024 -0.00013 -8.9e-05 -0.0008 -0.00079 -0.00023 -0.0066 -0.0013 -0.00075 -0.0014 -0.0029 0.017 -0.0029 -0.0067 -0.0015 -0.00037 -0.0002 -0.00087 -0.0024 -0.00083 -0.0017 -0.00073 0.0032 -0.00029 -0.0027 -0.00018 -0.00055 -0.0019 -0.0004 -0.00037 -0.0007 -0.00081 -0.00041 0.0041 0.0039 -0.00029 -0.001 -0.0022 -0.00088 -0.00056 0.026 -0.00058 -0.001 -0.00081 -0.00074 -0.00069 -0.0031 -0.00059 -0.00068 -0.00091 -0.0019 0.0045 -0.0029 -0.002 -0.00095 -0.00038 -0.00074 -0.00043 -0.0018 0.0043 -0.00097 -0.0021
interaction_26 0.0032 -0.0037 -0.00061 0.0086 -0.0065 -0.0032 -0.0039 0.013 -0.00019 -0.0024 -0.005 0.018 -0.0036 -0.0023 -0.0015 0.045 -0.008 -0.0015 -0.0012 -0.0046 -0.0038 -0.0012 -0.0026 -0.0021 -0.0031 -0.00095 -0.0016 -0.00069 -0.0015 -0.0007 -0.00013 1.0 -0.003 -0.0019 -0.0016 -0.0014 -0.00083 -0.00036 -0.00035 -0.0049 -0.0012 -0.00043 -0.002 -0.00015 -9.8e-05 -0.0022 -0.00033 -0.0024 -0.0011 -0.00026 -0.00088 -0.00047 -0.0054 -0.00028 -0.00027 -0.00014 -9.8e-05 -0.00088 -0.00087 -0.00025 -0.0072 -0.0014 -0.00082 -0.0015 -0.0031 -0.002 -0.0032 0.014 -0.0017 -0.0004 -0.00022 -0.00095 -0.0026 -0.00091 0.0047 -0.0008 0.0023 -0.00031 -0.0029 -0.0002 -0.0006 -0.0021 -0.00043 -0.00041 -0.00077 -0.00089 -0.00045 -0.0013 -0.0023 -0.00032 -0.0011 0.0027 -0.00097 -0.00062 0.023 -0.00064 -0.0011 -0.00088 -0.00081 -0.00076 -0.0015 -0.00064 -0.00074 -0.001 -0.002 0.0075 -0.0032 -0.0022 -0.0034 -0.00042 -0.00081 -0.00047 -0.002 -0.0022 -0.0011 -0.0023
interaction_27 0.012 -0.17 0.016 0.64 -0.14 -0.068 -0.081 -0.12 -0.0023 -0.051 -0.11 -0.094 -0.075 -0.049 -0.031 -0.025 0.38 -0.032 -0.026 -0.097 -0.079 -0.024 -0.054 -0.043 -0.065 -0.02 -0.034 -0.014 -0.031 -0.015 -0.0028 -0.003 1.0 -0.04 -0.033 -0.03 -0.017 -0.0075 -0.0074 -0.1 -0.026 -0.009 -0.041 -0.0032 -0.0021 -0.046 -0.007 -0.051 -0.023 -0.0054 -0.019 -0.0099 -0.11 -0.0058 -0.0056 -0.003 -0.0021 -0.018 -0.018 -0.0053 -0.15 -0.029 -0.017 -0.03 -0.0094 -0.16 0.016 0.34 0.021 -0.0085 -0.0045 -0.016 0.058 0.041 0.027 -0.006 0.084 -0.0026 -0.051 -0.0027 -0.011 -0.038 -0.0091 -0.0078 -0.015 0.0082 -0.0081 -0.019 -0.027 -0.0067 0.0012 -0.034 -0.02 -0.013 0.014 -0.011 -0.022 -0.018 -0.0019 0.0079 -0.024 -0.013 0.012 0.0054 0.17 -0.015 -0.067 -0.02 -0.018 -0.0087 -0.014 -0.0099 -0.041 0.1 -0.021 -0.049
interaction_28 -0.078 -0.0051 0.00096 0.041 -0.085 -0.043 -0.051 -0.077 -0.0025 0.6 -0.066 0.32 -0.047 -0.031 -0.02 -0.016 -0.1 0.26 -0.016 -0.061 -0.049 -0.015 -0.034 -0.027 -0.041 -0.012 -0.022 -0.0091 -0.019 -0.0092 -0.0017 -0.0019 -0.04 1.0 -0.021 -0.019 -0.011 -0.0047 -0.0046 -0.064 -0.016 -0.0057 -0.026 -0.002 -0.0013 -0.029 -0.0044 -0.032 -0.015 -0.0034 -0.012 -0.0062 -0.071 -0.0037 -0.0035 -0.0019 -0.0013 -0.012 -0.011 -0.0033 0.13 0.019 -0.0091 0.024 0.095 -0.1 -0.028 -0.096 -0.02 -0.0053 -0.00064 -0.013 -0.027 -0.011 -0.011 -0.0082 0.01 -0.0041 -0.038 -0.0026 -0.0039 -0.025 -0.0046 -0.003 -0.01 -0.0074 -0.0038 -0.01 -0.027 -0.0027 -0.0036 -0.017 -0.013 -0.0081 0.058 -0.0062 -0.014 -0.012 -0.0078 -0.0075 -0.039 -0.0077 -0.0079 -0.012 0.066 0.21 -0.042 -0.023 -0.025 -0.0055 -0.0095 -0.0062 -0.026 -0.025 -0.013 -0.031
interaction_30 -0.011 0.0037 -0.0076 -0.051 0.24 -0.035 -0.042 -0.064 -0.002 -0.026 -0.055 -0.049 -0.039 -0.025 -0.016 -0.013 -0.087 -0.017 -0.013 -0.05 -0.041 -0.013 -0.028 -0.023 -0.034 -0.01 -0.018 -0.0075 -0.016 -0.0076 -0.0014 -0.0016 -0.033 -0.021 1.0 -0.015 -0.009 -0.0039 -0.0038 -0.053 -0.014 -0.0047 -0.021 -0.0017 -0.0011 -0.024 -0.0036 -0.027 -0.012 -0.0028 -0.0096 -0.0052 -0.059 -0.003 -0.0029 -0.0015 -0.0011 -0.0096 -0.0094 -0.0028 -0.079 -0.015 0.14 0.14 0.093 -0.083 0.25 -0.08 -0.017 -0.0044 0.00027 -0.0098 0.026 -0.0061 0.023 -0.0066 0.013 -0.0034 0.033 -0.0022 -0.0037 -0.018 -0.0021 -0.0044 -0.0054 -0.0033 -0.0024 -0.0079 -0.013 -0.0017 -0.0078 0.056 -0.0023 -0.0067 0.0043 -0.0061 -0.011 0.013 -0.0061 -0.0061 -0.011 0.0096 -0.0058 -0.0097 0.055 -0.021 -0.032 -0.014 -0.017 -0.0045 0.0038 -0.004 -0.011 -0.013 -0.0052 0.042
interaction_33 0.011 -0.0039 -0.0021 -0.046 0.22 -0.032 -0.038 -0.057 -0.0018 -0.024 -0.049 -0.044 0.39 -0.023 -0.014 -0.012 -0.078 -0.015 -0.012 -0.045 -0.037 -0.011 -0.025 -0.02 -0.03 -0.0093 -0.016 -0.0067 -0.014 -0.0068 -0.0013 -0.0014 -0.03 -0.019 -0.015 1.0 -0.0081 -0.0035 -0.0034 -0.048 -0.012 -0.0042 -0.019 -0.0015 -0.00095 -0.021 -0.0032 -0.024 -0.011 -0.0025 -0.0086 -0.0046 -0.053 -0.0027 -0.0026 -0.0014 -0.00095 -0.0086 -0.0084 -0.0025 0.17 0.015 0.012 -0.012 -0.019 -0.074 -0.017 -0.072 -0.015 -0.0039 0.00082 -0.004 -0.011 -0.0061 0.099 -0.0054 0.012 -0.003 -0.022 -0.002 -0.0058 -0.019 -0.0042 0.0022 0.0016 -0.0044 -0.0044 0.0033 0.008 -0.0012 -0.0072 0.029 -0.0081 -0.0009 -0.0034 -0.0062 -0.0089 -0.0029 -0.004 -0.0049 0.026 0.0035 -0.0073 -0.0053 -0.01 0.02 -0.022 -0.01 -0.0098 -0.0041 -0.0048 0.00067 -0.016 -0.019 0.0027 -0.0012
interaction_34 0.023 0.013 0.00051 -0.027 0.052 0.11 -0.022 -0.033 -0.0011 -0.014 -0.029 -0.026 0.095 0.18 -0.0085 -0.0069 -0.046 -0.0088 -0.007 -0.026 -0.022 -0.0067 -0.015 -0.012 -0.018 -0.0055 -0.0094 -0.004 -0.0084 -0.004 -0.00076 -0.00083 -0.017 -0.011 -0.009 -0.0081 1.0 -0.0021 -0.002 -0.028 -0.0071 -0.0025 -0.011 -0.00088 -0.00056 -0.013 -0.0019 -0.014 -0.0064 -0.0015 -0.0051 -0.0027 -0.031 -0.0016 -0.0015 -0.00081 -0.00056 -0.005 -0.005 -0.0014 0.098 0.021 -0.0047 -0.0065 -0.0096 -0.044 -0.014 -0.042 -0.0084 -0.0023 -0.0012 -0.00099 -0.013 0.003 0.036 -0.0033 0.023 -0.0018 -0.015 -0.0011 -0.0034 -0.0091 -0.0025 -0.0023 0.0025 -0.0051 -0.0026 -0.0047 -0.005 -0.0018 -0.0044 0.055 0.037 -0.0018 0.0006 -0.0037 -0.0063 -0.0051 -0.0047 -0.003 0.012 -0.0037 -0.0043 -0.0025 -0.012 -0.0052 -0.016 0.0054 0.0048 -0.0024 -0.0046 -0.0027 -0.01 -0.011 -0.0061 -0.013
interaction_35 0.0027 -0.0018 -0.0013 -0.012 -0.0015 -0.0081 0.072 -0.015 -0.00046 -0.006 -0.012 -0.011 0.078 -0.0058 0.046 -0.003 -0.02 -0.0038 -0.003 -0.011 -0.0093 -0.0029 -0.0064 -0.0051 -0.0077 -0.0024 -0.0041 -0.0017 -0.0036 -0.0017 -0.00033 -0.00036 -0.0075 -0.0047 -0.0039 -0.0035 -0.0021 1.0 -0.00088 -0.012 -0.0031 -0.0011 -0.0049 -0.00038 -0.00024 -0.0055 -0.00083 -0.006 -0.0028 -0.00064 -0.0022 -0.0012 -0.013 -0.00069 -0.00066 -0.00035 -0.00024 -0.0022 -0.0021 -0.00063 -0.018 -0.0034 -0.002 -0.0037 -0.0078 0.047 -0.0079 -0.018 -0.0042 -0.001 -0.00053 0.0002 0.0082 -0.0023 0.019 -0.002 -0.0069 -0.00078 0.0033 -0.0005 -0.0015 -0.0053 -0.0011 -0.001 -0.0019 -0.0022 -0.0011 -0.0031 0.0086 -0.0008 -0.0027 0.0033 0.013 0.0024 -0.0031 -0.0016 -0.0027 0.037 0.01 -0.0019 -0.0046 -0.0016 -0.0019 -2.6e-05 -0.0051 -0.011 0.00023 -0.0053 -0.0037 -0.001 0.046 -0.0012 -1.3e-05 -0.00073 -0.00035 -0.0025
interaction_36 -0.0027 0.011 -0.0015 -0.011 0.019 -0.0079 -0.0094 0.024 -0.00045 -0.0059 -0.012 -0.011 0.045 -0.0057 -0.0036 0.14 -0.019 -0.0038 -0.003 -0.011 -0.0091 -0.0028 -0.0063 -0.005 -0.0075 -0.0023 -0.004 -0.0017 -0.0036 -0.0017 -0.00032 -0.00035 -0.0074 -0.0046 -0.0038 -0.0034 -0.002 -0.00088 1.0 -0.012 -0.003 -0.001 -0.0048 -0.00037 -0.00024 -0.0054 -0.00081 -0.0059 -0.0027 -0.00062 -0.0021 -0.0012 -0.013 -0.00068 -0.00065 -0.00034 -0.00024 -0.0021 -0.0021 -0.00061 -0.018 -0.0033 -0.002 -0.0037 -0.0025 0.0062 -0.0053 0.019 -0.0041 -0.00098 -0.00052 0.0003 0.0027 -0.0022 0.00081 -0.0019 -0.0067 -0.00076 0.019 -0.00049 -0.0015 0.002 0.0047 -0.00099 -0.0019 -0.0022 -0.0011 0.0009 2.5e-05 -0.00078 -0.0027 0.018 -0.0024 0.0025 0.001 -0.0016 -0.00037 0.00068 -0.002 -0.0019 0.0012 -0.0016 -0.0018 -0.0024 -0.005 -0.0072 -0.00027 -0.0052 -0.0034 -0.001 -0.002 -0.0012 0.0015 -0.0029 -0.00024 0.0043
interaction_37 -0.013 -0.01 -0.012 -0.16 0.76 -0.11 -0.13 -0.2 -0.0041 -0.082 -0.17 -0.15 -0.12 -0.079 -0.05 -0.04 0.61 -0.052 -0.041 -0.16 -0.13 -0.039 -0.087 -0.07 -0.1 -0.032 -0.055 -0.023 -0.049 -0.024 -0.0045 -0.0049 -0.1 -0.064 -0.053 -0.048 -0.028 -0.012 -0.012 1.0 -0.042 -0.015 -0.066 -0.0052 -0.0033 -0.074 -0.011 -0.082 -0.038 -0.0086 -0.03 -0.016 -0.18 -0.0094 -0.009 -0.0048 -0.0033 -0.03 -0.029 -0.0085 -0.24 -0.046 -0.028 -0.05 0.02 -0.26 -0.051 0.56 -0.044 -0.0069 -0.0043 -0.024 0.068 -0.013 0.049 0.008 -0.013 -0.0072 -0.062 -0.0026 -0.015 -0.059 -0.013 0.00059 -0.0087 -0.016 -0.0097 -0.0093 -0.0061 0.0099 -0.027 -0.0061 -8.5e-05 -0.0032 -0.015 -0.02 -0.032 0.035 -0.02 -0.018 0.037 0.0036 0.018 -0.014 -0.022 0.016 -0.061 -0.0042 0.08 -0.004 0.016 -0.012 -0.055 -0.049 -0.0012 0.21
interaction_38 -0.0046 0.0022 -0.003 -0.04 0.083 -0.028 -0.033 -0.05 -0.0016 0.21 -0.043 -0.039 0.13 -0.02 -0.013 -0.01 -0.068 0.44 -0.011 -0.04 -0.032 -0.0099 -0.022 -0.018 -0.026 -0.0081 -0.014 -0.0059 -0.013 -0.006 -0.0011 -0.0012 -0.026 -0.016 -0.014 -0.012 -0.0071 -0.0031 -0.003 -0.042 1.0 -0.0037 -0.017 -0.0013 -0.00084 -0.019 -0.0029 -0.021 -0.0095 -0.0022 -0.0076 -0.0041 -0.046 -0.0024 -0.0023 -0.0012 -0.00084 -0.0075 -0.0074 -0.0022 0.085 -0.0076 -0.007 -0.011 0.099 -0.065 -0.0088 -0.063 -0.014 -0.0035 -0.0018 -0.0059 -0.02 -0.007 0.066 -0.0015 0.012 -0.0027 -0.018 -0.0017 -0.0051 -0.017 -0.0037 1.1e-05 -0.0056 -0.006 -0.0039 0.0086 -0.018 -0.0027 -0.0055 -0.0069 -0.0068 -0.0053 0.056 -0.0055 -0.0087 -0.0051 -0.0061 -0.0037 -0.027 0.0011 -0.0064 -0.0085 -0.0093 0.095 -0.026 -0.0099 0.0036 -0.0018 -0.0052 -0.0041 -0.016 -0.016 -0.0091 -0.02
interaction_40 0.011 0.0075 -0.0021 -0.014 -0.019 0.13 -0.012 -0.017 -0.00056 -0.0072 -0.015 -0.013 -0.011 -0.007 -0.0044 -0.0036 -0.024 -0.0046 -0.0036 -0.014 -0.011 -0.0035 -0.0077 -0.0062 -0.0092 -0.0028 -0.0049 -0.0021 -0.0044 -0.0021 -0.00039 -0.00043 -0.009 -0.0057 -0.0047 -0.0042 -0.0025 -0.0011 -0.001 -0.015 -0.0037 1.0 -0.0059 -0.00046 -0.00029 -0.0065 -0.00099 -0.0072 -0.0033 -0.00076 -0.0026 -0.0014 -0.016 -0.00083 -0.0008 -0.00042 -0.00029 -0.0026 -0.0026 -0.00075 -0.022 -0.0041 0.005 0.019 -0.0031 -0.023 0.11 -0.022 -0.005 -0.0012 -0.00064 -0.0028 -0.0077 -0.0027 0.0032 -0.0024 0.0015 -0.00093 -0.0087 -0.0006 -0.0018 -0.0053 -0.0013 -0.0012 0.00038 -0.0026 -0.0013 -0.0038 0.012 -0.00095 -0.0014 0.029 -0.0029 -0.0018 -0.0004 -0.0019 -0.0032 -0.0026 -0.0024 -0.0023 -0.0024 0.0013 -0.0022 -0.003 -0.003 0.0065 -0.0074 0.016 0.005 -0.0012 0.0077 -0.0014 -0.006 -0.0045 -0.0032 -0.0069
interaction_44 0.0028 0.0035 0.0062 -0.064 -0.088 0.61 -0.053 -0.079 -0.0025 -0.033 -0.068 -0.061 -0.049 0.84 -0.02 -0.016 -0.11 -0.021 -0.017 -0.063 -0.051 -0.016 -0.035 -0.028 -0.042 -0.013 -0.022 -0.0094 -0.02 -0.0095 -0.0018 -0.002 -0.041 -0.026 -0.021 -0.019 -0.011 -0.0049 -0.0048 -0.066 -0.017 -0.0059 1.0 -0.0021 -0.0013 -0.03 -0.0045 -0.033 -0.015 -0.0035 -0.012 -0.0065 -0.074 -0.0038 -0.0036 -0.0019 -0.0013 -0.012 -0.012 -0.0034 0.26 -0.0039 -0.011 -0.02 -0.036 -0.1 -0.034 -0.1 -0.023 0.0046 -0.0029 -0.0019 -0.035 0.0013 -0.013 -0.0012 -0.019 -0.0043 -0.033 -0.0027 -0.0081 0.0047 -0.0059 -0.0044 0.032 -0.0007 -0.0051 0.0083 0.051 -0.003 -0.012 0.037 -0.012 -0.0084 -0.0036 -0.008 -0.015 -0.012 -0.0088 -0.0037 0.061 -0.0088 -0.01 -0.0085 -0.027 0.048 -0.041 0.033 0.024 -0.0057 0.0037 -0.0026 -0.022 -0.011 -0.014 -0.032
interaction_45 -0.002 0.0038 -0.00055 -0.005 -0.0068 0.0097 0.028 -0.0062 -0.0002 -0.0026 -0.0053 -0.0048 -0.0038 0.048 0.026 -0.0013 -0.0085 -0.0016 -0.0013 -0.0049 -0.004 -0.0012 -0.0027 -0.0022 -0.0033 -0.001 -0.0017 -0.00073 -0.0016 -0.00074 -0.00014 -0.00015 -0.0032 -0.002 -0.0017 -0.0015 -0.00088 -0.00038 -0.00037 -0.0052 -0.0013 -0.00046 -0.0021 1.0 -0.0001 -0.0023 -0.00035 -0.0026 -0.0012 -0.00027 -0.00094 -0.0005 -0.0057 -0.0003 -0.00028 -0.00015 -0.0001 -0.00093 -0.00092 -0.00027 -0.0077 -0.0015 -0.00087 -0.0016 -0.0014 0.019 -0.0034 -0.0078 -0.0018 -0.00043 -0.00023 -0.001 -0.0028 -0.00096 0.0073 -0.00085 -0.0036 -0.00033 0.0072 -0.00021 -0.00063 -0.0023 -0.00046 -0.00043 -0.00081 -0.00094 -0.00048 -0.0013 -0.0024 -0.00034 -0.0012 0.0096 0.0049 -0.00065 -0.0013 -0.00068 -0.0012 -0.00094 -0.00086 -0.00081 0.0054 -0.00068 -0.00079 0.0047 -0.0022 -0.0028 -0.0015 -0.0023 -0.0036 -0.00044 -0.00086 -0.0005 -0.0021 0.0086 -0.0011 -0.0025
interaction_46 -0.00013 0.0025 -6.1e-06 -0.0032 -0.0044 0.0096 -0.0026 0.0092 -0.00013 -0.0016 -0.0034 -0.0031 -0.0024 0.026 -0.001 0.029 -0.0054 -0.001 -0.00083 -0.0031 -0.0025 -0.00079 -0.0017 -0.0014 -0.0021 -0.00064 -0.0011 -0.00047 -0.00099 -0.00047 -8.9e-05 -9.8e-05 -0.0021 -0.0013 -0.0011 -0.00095 -0.00056 -0.00024 -0.00024 -0.0033 -0.00084 -0.00029 -0.0013 -0.0001 1.0 -0.0015 -0.00023 -0.0016 -0.00075 -0.00017 -0.0006 -0.00032 -0.0037 -0.00019 -0.00018 -9.6e-05 -6.6e-05 -0.0006 -0.00059 -0.00017 -0.0049 -0.00093 -0.00056 -0.001 -0.0021 0.0031 -0.0022 0.005 -0.0011 -0.00027 -0.00015 -0.00064 -0.0018 -0.00061 -0.0013 0.011 -0.0023 -0.00021 0.0012 -0.00014 -0.0004 -0.0014 -0.00029 -0.00028 -0.00052 -0.0006 -0.00031 -0.00085 0.0025 -0.00022 -0.00074 0.006 -0.00065 -0.00042 -0.00084 -0.00043 -0.00074 -0.0006 -0.00055 -0.00051 0.00053 -0.00044 -0.0005 -0.00068 -0.0014 0.0027 -0.0022 -0.0015 0.0034 -0.00028 -0.00055 -0.00032 -0.0014 -0.0015 -0.00072 -0.0016
interaction_47 0.0032 0.046 0.0039 -0.072 -0.098 0.68 -0.059 -0.089 -0.0028 -0.037 -0.076 -0.069 -0.055 -0.036 -0.023 -0.018 0.28 -0.023 -0.019 -0.07 -0.057 -0.018 -0.039 -0.032 -0.047 -0.014 -0.025 -0.011 -0.022 -0.011 -0.002 -0.0022 -0.046 -0.029 -0.024 -0.021 -0.013 -0.0055 -0.0054 -0.074 -0.019 -0.0065 -0.03 -0.0023 -0.0015 1.0 -0.0051 -0.037 -0.017 -0.0039 -0.013 -0.0072 -0.082 -0.0042 -0.0041 -0.0022 -0.0015 -0.013 -0.013 -0.0038 -0.11 -0.021 -0.013 -0.023 -0.046 -0.12 -0.038 0.29 -0.025 -0.0031 -0.0033 -0.0076 -0.03 -0.011 0.014 -0.0086 -0.027 -0.0048 -0.04 -0.003 -0.0091 -0.023 -0.0056 -0.0062 0.014 -0.0098 -0.0069 -0.009 0.1 -0.0036 -0.0084 -0.024 0.0087 0.021 -0.011 -0.0097 -0.016 -0.012 -0.0088 -0.01 0.076 -0.0098 -0.0097 -0.014 -0.03 0.015 0.021 0.035 0.012 -0.0053 0.014 -0.0072 -0.031 0.035 -0.005 -0.032
interaction_48 0.018 -0.0041 0.00084 -0.011 -0.015 0.045 -0.0089 -0.013 -0.00043 0.07 -0.012 -0.01 -0.0083 0.073 -0.0034 -0.0028 -0.018 0.1 -0.0028 -0.011 -0.0086 -0.0027 -0.0059 -0.0048 -0.0071 -0.0022 -0.0038 -0.0016 -0.0034 -0.0016 -0.0003 -0.00033 -0.007 -0.0044 -0.0036 -0.0032 -0.0019 -0.00083 -0.00081 -0.011 -0.0029 -0.00099 -0.0045 -0.00035 -0.00023 -0.0051 1.0 -0.0056 -0.0026 -0.00059 -0.002 -0.0011 -0.012 -0.00064 -0.00062 -0.00033 -0.00023 -0.002 -0.002 -0.00058 0.039 -0.0012 -0.0019 -0.0035 0.00087 -0.017 -0.0065 -0.017 -0.0039 -0.00093 -0.0005 -0.0022 -0.006 -0.0021 0.014 -0.0018 0.021 -0.00072 -0.0048 -0.00046 -0.0014 -0.0011 -0.001 -0.00094 0.0017 -0.002 -0.001 -0.0029 0.0089 -0.00074 -0.0025 -3.5e-06 -0.0022 -0.0014 0.0014 -0.0015 -0.0025 -0.002 -0.0019 -0.0018 -0.0053 -0.0015 -0.0017 -0.0023 -0.0047 0.0028 -0.0073 7.5e-05 0.0074 -0.00096 0.0014 -0.0011 -0.0046 -0.0012 -0.0025 -0.0042
interaction_50 0.025 0.054 -0.011 -0.079 -0.11 -0.055 0.63 -0.098 -0.0031 -0.041 -0.084 -0.076 -0.061 -0.039 -0.025 -0.02 -0.13 -0.026 -0.021 -0.078 -0.063 -0.02 -0.043 -0.035 -0.052 -0.016 -0.028 -0.012 -0.025 -0.012 -0.0022 -0.0024 -0.051 -0.032 -0.027 -0.024 -0.014 -0.006 -0.0059 -0.082 -0.021 -0.0072 -0.033 -0.0026 -0.0016 -0.037 -0.0056 1.0 -0.019 -0.0043 -0.015 -0.008 -0.091 -0.0047 -0.0045 -0.0024 -0.0016 -0.015 -0.015 -0.0042 -0.12 -0.023 -0.014 -0.025 -0.052 0.32 -0.053 -0.12 -0.027 -0.0059 0.0085 0.033 -0.025 -0.0086 -0.0092 -0.011 -0.03 -0.0053 -0.015 -0.0034 -0.01 -0.012 0.01 -0.0041 0.0008 0.0021 -0.0034 0.022 0.047 0.0039 0.0011 -0.02 0.004 0.022 -0.014 -0.0084 -0.011 0.0026 -0.0049 -0.0059 0.019 -0.005 -0.012 -0.0021 -0.03 -0.068 0.2 -0.031 -0.045 -0.007 0.021 0.0023 0.05 -0.0092 0.063 -0.019
interaction_55 -0.015 0.041 -0.0032 -0.036 -0.05 -0.025 0.29 -0.045 -0.0014 -0.019 -0.039 -0.035 -0.028 -0.018 0.75 -0.0092 -0.061 -0.012 -0.0094 -0.036 -0.029 -0.0089 -0.02 -0.016 -0.024 -0.0073 -0.013 -0.0053 -0.011 -0.0054 -0.001 -0.0011 -0.023 -0.015 -0.012 -0.011 -0.0064 -0.0028 -0.0027 -0.038 -0.0095 -0.0033 -0.015 -0.0012 -0.00075 -0.017 -0.0026 -0.019 1.0 -0.002 -0.0068 -0.0036 -0.042 -0.0021 -0.0021 -0.0011 -0.00075 -0.0068 -0.0067 -0.0019 -0.056 -0.011 -0.0063 -0.012 -0.024 0.15 -0.025 -0.056 -0.012 -0.0031 0.0057 0.0077 -0.019 -0.0052 -0.0033 -0.0062 -0.021 -0.0024 0.097 -0.0015 -0.0033 -0.013 -0.0033 0.0066 0.023 0.036 -0.0035 0.0042 0.0042 -3.2e-06 -0.0041 -0.016 0.0033 0.0056 -0.0089 -0.0012 -0.00036 0.0058 -0.0033 -0.0058 0.017 -0.0037 -0.0057 0.015 -0.012 -0.033 0.0036 -0.012 -0.017 -0.0032 -0.0023 -0.0037 0.00076 0.0034 0.014 0.014
interaction_56 -0.0068 0.0064 -0.0012 -0.0083 -0.011 -0.0057 0.031 0.016 -0.00033 -0.0043 -0.0089 -0.008 -0.0064 -0.0041 0.081 0.11 -0.014 -0.0027 -0.0022 -0.0082 -0.0066 -0.0021 -0.0046 -0.0037 -0.0055 -0.0017 -0.0029 -0.0012 -0.0026 -0.0012 -0.00023 -0.00026 -0.0054 -0.0034 -0.0028 -0.0025 -0.0015 -0.00064 -0.00062 -0.0086 -0.0022 -0.00076 -0.0035 -0.00027 -0.00017 -0.0039 -0.00059 -0.0043 -0.002 1.0 -0.0016 -0.00084 -0.0096 -0.00049 -0.00047 -0.00025 -0.00017 -0.0016 -0.0015 -0.00045 -0.013 -0.0024 -0.0015 -0.0027 -0.0055 0.022 -0.0057 -0.0015 -0.003 -0.00071 -0.00038 -0.0017 -0.0046 -0.0016 -0.0033 -0.0014 -0.006 -0.00055 0.05 -0.00035 -0.0011 -0.0038 -0.00077 -0.00072 -0.0014 -0.0016 -0.0008 0.006 -0.0025 -0.00057 -0.0019 -0.0028 -0.0017 -0.0011 -0.0022 -0.0011 0.011 -0.0016 -0.0014 -0.0013 -0.0038 -0.0011 -0.0013 -0.0018 -0.0036 -0.009 0.0012 -0.0022 -0.0049 -0.00074 -0.0014 0.0064 0.0034 -0.0022 0.0046 -0.0041
interaction_57 0.005 0.03 -0.0035 -0.029 -0.039 -0.02 0.23 -0.036 -0.0011 -0.015 -0.031 -0.028 -0.022 -0.014 -0.0091 -0.0073 0.11 -0.0094 -0.0075 -0.028 -0.023 -0.0071 -0.016 -0.013 -0.019 -0.0058 -0.01 -0.0042 -0.0089 -0.0043 -0.00081 -0.00088 -0.019 -0.012 -0.0096 -0.0086 -0.0051 -0.0022 -0.0021 -0.03 -0.0076 -0.0026 -0.012 -0.00094 -0.0006 -0.013 -0.002 -0.015 -0.0068 -0.0016 1.0 -0.0029 -0.033 -0.0017 -0.0016 -0.00086 -0.0006 -0.0054 -0.0053 -0.0015 -0.044 -0.0084 -0.005 -0.0092 -0.019 -0.046 -0.019 0.12 -0.0097 -0.0025 -0.0013 0.0099 -0.011 -0.0033 0.021 -0.0024 -0.013 -0.0019 0.014 -0.0012 -0.0036 -0.0091 -0.0026 0.017 0.0096 0.014 -0.00055 0.0019 0.012 -0.002 0.007 -0.0085 -0.0018 0.011 -0.0036 0.00077 0.0098 0.013 0.0098 0.00061 0.0045 -0.00083 -0.0045 0.0049 -0.01 -0.026 0.043 -0.0055 -0.015 0.0046 0.0012 0.0034 0.002 0.0015 0.025 -0.0098
interaction_58 0.011 0.019 -0.0021 -0.015 -0.021 -0.011 0.11 -0.019 -0.00061 0.0055 -0.016 -0.015 -0.012 -0.0077 0.017 -0.0039 -0.026 0.29 -0.004 -0.015 -0.012 -0.0038 -0.0084 -0.0068 -0.01 -0.0031 -0.0054 -0.0023 -0.0048 -0.0023 -0.00043 -0.00047 -0.0099 -0.0062 -0.0052 -0.0046 -0.0027 -0.0012 -0.0012 -0.016 -0.0041 -0.0014 -0.0065 -0.0005 -0.00032 -0.0072 -0.0011 -0.008 -0.0036 -0.00084 -0.0029 1.0 -0.018 -0.00091 -0.00088 -0.00046 -0.00032 -0.0029 -0.0028 -0.00083 -0.024 -0.0045 -0.0027 -0.0049 -0.01 0.062 -0.01 -0.024 -0.0055 -0.0013 -0.0007 -0.0031 -0.0085 -0.003 0.065 -0.0026 -0.0024 -0.001 0.0071 -0.00066 -0.002 -0.007 -0.0014 0.012 -0.0025 -0.0029 0.0026 0.0033 -0.0033 -0.0011 0.0049 -0.0064 -0.0012 0.00098 0.0064 -0.0021 0.0049 -0.00079 -0.00038 -0.0025 -0.0076 -0.0021 4.7e-05 0.00046 -0.0039 -0.011 0.03 -0.00086 -0.0081 -0.0014 -0.0026 -0.0016 -0.0056 -0.0071 0.0017 -0.0068
interaction_60 -0.033 0.17 -0.026 -0.18 -0.24 -0.12 -0.14 0.93 -0.007 -0.091 -0.19 -0.17 -0.13 -0.088 -0.056 -0.045 -0.3 -0.058 -0.046 -0.17 -0.14 -0.043 -0.096 -0.078 -0.12 -0.036 -0.061 -0.026 -0.055 -0.026 -0.0049 -0.0054 -0.11 -0.071 -0.059 -0.053 -0.031 -0.013 -0.013 -0.18 -0.046 -0.016 -0.074 -0.0057 -0.0037 -0.082 -0.012 -0.091 -0.042 -0.0096 -0.033 -0.018 1.0 -0.01 -0.01 -0.0053 -0.0037 -0.033 -0.032 -0.0095 -0.27 -0.051 -0.031 -0.056 -0.12 0.71 -0.12 -0.28 -0.047 0.044 0.014 0.067 -0.072 -0.0084 -0.04 -0.015 -0.079 -0.0024 0.078 0.0012 -0.019 0.047 0.028 0.0081 0.038 -0.0072 0.0059 0.0067 0.011 -0.0011 -0.00085 -0.025 0.027 0.012 -0.022 0.042 0.087 -0.0072 0.081 0.003 0.045 -0.014 -0.015 0.046 -0.071 -0.14 0.21 -0.06 -0.078 0.011 -0.0014 0.028 0.18 -0.047 0.036 -0.035
interaction_66 -0.0083 0.0054 -0.0013 -0.0091 -0.012 -0.0063 -0.0075 0.048 -0.00036 -0.0047 -0.0097 -0.0087 -0.0069 -0.0045 -0.0029 0.23 -0.015 -0.003 -0.0024 -0.0089 -0.0072 -0.0022 -0.005 -0.004 -0.006 -0.0018 -0.0032 -0.0013 -0.0028 -0.0013 -0.00025 -0.00028 -0.0058 -0.0037 -0.003 -0.0027 -0.0016 -0.00069 -0.00068 -0.0094 -0.0024 -0.00083 -0.0038 -0.0003 -0.00019 -0.0042 -0.00064 -0.0047 -0.0021 -0.00049 -0.0017 -0.00091 -0.01 1.0 -0.00052 -0.00027 -0.00019 -0.0017 -0.0017 -0.00049 -0.014 -0.0026 -0.0016 -0.0008 -0.006 0.036 -0.0062 -0.014 -0.0032 -0.00078 -0.00041 -0.0018 -0.005 -0.0017 -0.0036 -0.0015 -0.0066 0.0094 0.0079 -0.00039 -0.0012 -0.0041 0.0064 -0.00078 -0.0015 -0.0017 -0.00087 0.01 0.0012 -0.00062 -0.0021 0.00066 -0.0019 0.0039 -0.0024 -0.0012 0.027 -0.0017 -0.0016 -0.0015 0.0044 -0.0012 -0.0014 -0.0019 -0.0039 -0.0098 0.015 -0.0041 -0.0065 -0.0008 -0.0016 0.019 0.012 0.0018 -0.0021 -0.0031
interaction_68 -0.0022 0.0027 -0.0012 -0.0087 -0.012 -0.006 -0.0072 0.037 -0.00035 0.014 -0.0093 -0.0083 -0.0067 -0.0043 -0.0027 0.034 -0.015 0.15 -0.0023 -0.0085 -0.0069 -0.0021 -0.0048 -0.0038 -0.0057 -0.0018 -0.003 -0.0013 -0.0027 -0.0013 -0.00024 -0.00027 -0.0056 -0.0035 -0.0029 -0.0026 -0.0015 -0.00066 -0.00065 -0.009 -0.0023 -0.0008 -0.0036 -0.00028 -0.00018 -0.0041 -0.00062 -0.0045 -0.0021 -0.00047 -0.0016 -0.00088 -0.01 -0.00052 1.0 -0.00026 -0.00018 -0.0016 -0.0016 -0.00047 -0.013 -0.0025 -0.0015 -0.0028 -0.0047 0.022 -0.0059 -0.00078 -0.0031 -0.00075 -0.0004 -0.0018 -0.0048 -0.0017 0.021 -0.0015 -0.0053 -0.00058 0.016 -0.00037 -0.0011 -0.0039 -0.0008 -0.00075 -0.0014 -0.0016 0.0064 -0.0023 -0.0042 -0.00059 0.028 -0.0017 -0.0018 -0.0011 -0.0023 -0.0012 0.00099 -0.0016 -0.0015 -0.0014 -0.0063 -0.0012 -0.0014 -0.0018 -0.0021 -0.0056 0.015 -0.00085 -0.0041 0.015 -0.0015 -0.00088 -0.002 -0.0024 -0.002 -0.0014
interaction_70 0.0027 0.0077 -0.00068 -0.0046 -0.0063 -0.0032 -0.0038 -0.0057 0.76 -0.0024 -0.0049 -0.0044 -0.0035 -0.0023 -0.0015 -0.0012 -0.0078 -0.0015 -0.0012 -0.0045 -0.0037 -0.0011 -0.0025 -0.002 -0.003 -0.00093 -0.0016 -0.00068 -0.0014 -0.00068 -0.00013 -0.00014 -0.003 -0.0019 -0.0015 -0.0014 -0.00081 -0.00035 -0.00034 -0.0048 -0.0012 -0.00042 -0.0019 -0.00015 -9.6e-05 -0.0022 -0.00033 -0.0024 -0.0011 -0.00025 -0.00086 -0.00046 -0.0053 -0.00027 -0.00026 1.0 -9.6e-05 -0.00086 -0.00085 -0.00025 -0.0071 -0.0013 -0.00081 0.0027 -0.0031 -0.0074 0.042 -0.0072 -0.0016 -0.0004 -0.00021 -0.00093 -6.4e-05 -0.00089 0.0015 -0.00078 -0.0033 -0.00031 -0.00064 -0.0002 -0.00058 -0.0021 -0.00042 -0.0004 -0.00075 -0.00087 -0.00044 0.0037 -0.0022 -0.00031 -0.0011 0.027 -0.00095 -0.0006 -0.0012 -0.00063 -0.0011 -0.00086 -0.0008 -0.00074 -0.0033 0.019 -0.00073 0.0052 -0.002 -0.005 -0.0031 0.0038 -0.0013 -0.00041 -0.00079 -0.00046 0.0043 -0.0021 -0.001 -0.0023
interaction_77 -0.0031 -0.0057 -0.00047 -0.0032 -0.0044 -0.0022 -0.0026 -0.004 0.52 -0.0016 -0.0034 -0.0031 -0.0024 -0.0016 -0.001 -0.00081 0.012 -0.001 -0.00083 -0.0031 -0.0025 -0.00079 -0.0017 -0.0014 -0.0021 -0.00064 -0.0011 -0.00047 -0.00099 -0.00047 -8.9e-05 -9.8e-05 -0.0021 -0.0013 -0.0011 -0.00095 -0.00056 -0.00024 -0.00024 -0.0033 -0.00084 -0.00029 -0.0013 -0.0001 -6.6e-05 -0.0015 -0.00023 -0.0016 -0.00075 -0.00017 -0.0006 -0.00032 -0.0037 -0.00019 -0.00018 -9.6e-05 1.0 -0.0006 -0.00059 -0.00017 -0.0049 -0.00093 -0.00056 -0.001 -0.0021 -0.0051 0.031 -0.005 -0.0011 -0.00027 -0.00015 -0.00064 -0.0018 -0.00061 -0.0013 -0.00054 -0.0023 -0.00021 -0.002 -0.00014 -0.0004 -0.0014 -0.00029 -0.00028 -0.00052 -0.0006 -0.00031 -0.00085 -0.0016 -0.00022 -0.00074 0.006 -0.00065 -0.00042 -0.00084 -0.00043 -0.00074 -0.0006 -0.00055 -0.00051 -0.0023 -0.00044 -0.0005 -0.00068 0.021 -0.0014 -0.0022 -0.0015 0.0063 -0.00028 -0.00055 -0.00032 -0.0014 -0.0015 -0.00072 -0.0016
interaction_78 -0.0074 -0.032 0.011 -0.029 -0.039 -0.02 -0.024 -0.036 -0.0011 0.36 -0.031 -0.027 -0.022 -0.014 -0.009 -0.0073 0.11 -0.0094 -0.0075 -0.028 -0.023 -0.0071 -0.016 -0.013 -0.019 -0.0058 -0.01 -0.0042 -0.0089 -0.0043 -0.0008 -0.00088 -0.018 -0.012 -0.0096 -0.0086 -0.005 -0.0022 -0.0021 -0.03 -0.0075 -0.0026 -0.012 -0.00093 -0.0006 -0.013 -0.002 -0.015 -0.0068 -0.0016 -0.0054 -0.0029 -0.033 -0.0017 -0.0016 -0.00086 -0.0006 1.0 -0.0053 -0.0015 -0.044 -0.0084 -0.005 -0.0085 0.059 -0.046 -0.011 0.072 -0.0096 -0.0025 -0.0013 -0.0047 -0.011 -0.0055 0.029 0.011 0.0021 -0.0019 -0.016 -0.0012 0.057 -0.0043 -0.0026 -2.9e-05 -0.0034 -0.0031 0.0017 -0.0013 -0.0045 0.0012 0.008 -0.0016 -0.0059 -0.0037 0.0022 -0.00076 -0.002 -0.002 0.015 -0.0046 -0.019 0.0039 0.0062 -0.0061 0.022 0.032 -0.015 0.023 -0.0089 -0.0025 3.4e-05 -0.0029 -0.011 -0.0017 0.003 -0.013
interaction_80 -0.021 -0.045 -0.0041 -0.028 -0.039 -0.019 -0.023 -0.035 -0.0011 0.36 -0.03 -0.027 -0.022 -0.014 -0.0089 -0.0072 -0.048 -0.0092 -0.0073 -0.028 -0.022 -0.0069 -0.015 -0.012 -0.018 -0.0057 -0.0098 -0.0041 -0.0088 -0.0042 -0.00079 -0.00087 -0.018 -0.011 -0.0094 -0.0084 -0.005 -0.0021 -0.0021 -0.029 -0.0074 -0.0026 -0.012 -0.00092 -0.00059 -0.013 -0.002 -0.015 -0.0067 -0.0015 -0.0053 -0.0028 -0.032 -0.0017 -0.0016 -0.00085 -0.00059 -0.0053 1.0 -0.0015 -0.043 -0.0082 0.16 0.2 -0.0097 -0.045 0.11 -0.044 -0.0064 -0.0024 -0.0013 -0.0035 -0.014 0.0025 0.024 0.0003 0.017 0.0014 -0.017 -0.0012 0.015 -0.012 -0.0026 0.0025 -0.0046 -0.0018 0.0018 -0.0043 -0.013 0.0012 -0.0047 -0.0085 -0.0058 -0.0037 0.025 -0.0022 -0.0056 -0.003 -0.0036 0.0021 -0.02 -0.00068 0.001 -0.0039 0.14 0.014 -0.019 0.0027 -0.012 -0.0025 -0.0048 -0.0028 -0.011 -0.0095 -0.0064 -0.014
interaction_88 0.023 -0.022 0.0034 -0.0082 -0.011 -0.0057 -0.0067 -0.01 -0.00033 0.1 -0.0088 -0.0079 -0.0063 -0.0041 -0.0026 -0.0021 -0.014 0.16 -0.0021 -0.0081 -0.0066 -0.002 -0.0045 -0.0036 -0.0054 -0.0017 -0.0029 -0.0012 -0.0026 -0.0012 -0.00023 -0.00025 -0.0053 -0.0033 -0.0028 -0.0025 -0.0014 -0.00063 -0.00061 -0.0085 -0.0022 -0.00075 -0.0034 -0.00027 -0.00017 -0.0038 -0.00058 -0.0042 -0.0019 -0.00045 -0.0015 -0.00083 -0.0095 -0.00049 -0.00047 -0.00025 -0.00017 -0.0015 -0.0015 1.0 0.017 -0.0024 -0.0014 0.0067 -0.00073 -0.013 0.016 -0.013 -0.0029 -0.00071 -0.00038 -0.0017 -0.0017 -0.0016 0.0099 0.0029 0.037 -0.00055 -0.0051 -0.00035 -0.001 -0.002 -0.00076 -0.00071 -0.0013 -0.0015 -0.00079 0.00057 -0.004 -0.00056 0.0076 -0.0028 -0.0017 -0.0011 0.0091 -0.0011 0.0045 -0.0015 -0.0014 -0.0013 -0.0048 -0.0011 -0.0013 -0.0017 -0.0036 -0.0041 -0.0056 -0.0021 -0.0014 -0.00073 -0.0014 -0.00083 -0.0035 -0.0038 -0.0019 -0.0041
event_type_Battle-No change of territory 0.0072 -0.11 0.042 0.047 -0.038 0.13 -0.19 -0.29 -0.0093 0.078 0.21 0.4 0.31 0.29 -0.074 -0.06 -0.4 0.11 0.13 0.43 0.33 0.14 -0.13 -0.1 -0.15 0.075 -0.082 0.071 0.13 0.077 -0.0066 -0.0072 -0.15 0.13 -0.079 0.17 0.098 -0.018 -0.018 -0.24 0.085 -0.022 0.26 -0.0077 -0.0049 -0.11 0.039 -0.12 -0.056 -0.013 -0.044 -0.024 -0.27 -0.014 -0.013 -0.0071 -0.0049 -0.044 -0.043 0.017 1.0 -0.069 -0.041 -0.075 -0.16 -0.38 -0.16 -0.37 0.11 -0.015 -0.0092 -0.031 -0.014 0.012 -0.003 0.017 0.049 -0.0025 -0.054 -0.0043 0.025 0.029 -0.021 -0.015 -0.017 -0.014 0.012 -0.0033 -0.022 -0.0012 -0.00037 0.018 -0.015 -0.028 0.015 -0.026 -0.052 -0.011 -0.027 0.012 -0.014 0.019 -0.0024 -0.011 -0.048 0.2 -0.15 0.058 -0.0056 -0.02 -0.025 -0.02 -0.076 0.028 -0.048 -0.11
event_type_Battle-Non-state actor overtakes territory 0.054 -0.078 0.0097 0.08 -0.018 -0.011 -0.037 -0.055 -0.0018 0.0082 0.077 0.11 0.022 0.0018 -0.014 -0.011 -0.076 -0.007 0.044 0.15 0.0037 -0.0026 -0.024 -0.02 -0.029 0.0012 -0.016 0.056 0.097 0.026 -0.0013 -0.0014 -0.029 0.019 -0.015 0.015 0.021 -0.0034 -0.0033 -0.046 -0.0076 -0.0041 -0.0039 -0.0015 -0.00093 -0.021 -0.0012 -0.023 -0.011 -0.0024 -0.0084 -0.0045 -0.051 -0.0026 -0.0025 -0.0013 -0.00093 -0.0084 -0.0082 -0.0024 -0.069 1.0 -0.0078 -0.014 -0.03 -0.072 -0.03 -0.07 -0.012 -0.0038 -0.002 -0.009 -0.02 -0.0072 0.0083 0.021 0.064 -0.003 -0.027 -0.0019 0.0019 -0.011 -0.0041 -0.0039 -0.0073 0.00036 0.0071 0.022 -0.021 -0.001 0.074 0.04 -0.0092 -0.0059 0.017 -0.0041 -0.01 -0.0062 -0.0069 0.037 -0.027 0.00091 -0.0071 -0.0088 0.035 -0.0079 -0.03 0.018 0.028 -0.0039 -0.0077 -0.0045 -0.019 -0.017 -0.01 -0.022
event_type_Headquarters or base established -0.032 -0.079 -0.0039 0.056 0.017 -0.017 -0.022 -0.033 -0.0011 0.048 -0.027 -0.023 -0.012 -0.013 -0.0084 -0.0068 -0.045 -0.0066 -0.007 -0.024 -0.021 -0.0066 -0.015 -0.012 -0.018 -0.0031 0.2 -0.0024 -0.0061 -0.004 -0.00075 -0.00082 -0.017 -0.0091 0.14 0.012 -0.0047 -0.002 -0.002 -0.028 -0.007 0.005 -0.011 -0.00087 -0.00056 -0.013 -0.0019 -0.014 -0.0063 -0.0015 -0.005 -0.0027 -0.031 -0.0016 -0.0015 -0.00081 -0.00056 -0.005 0.16 -0.0014 -0.041 -0.0078 1.0 -0.0086 -0.018 -0.043 -0.018 -0.042 -0.0083 -0.0023 -0.0012 -0.0054 -0.0088 -0.004 -0.0072 -0.0046 0.017 -0.0018 -0.017 -0.0011 0.00018 -0.011 -0.0025 -0.0023 -0.003 -0.005 0.0022 -0.0072 -0.011 -0.0018 -0.0013 -0.0097 -0.0055 -0.0035 -0.0019 -0.0036 -0.0062 -0.0026 -0.0046 -0.0043 -0.016 -0.00033 -0.0028 -0.0046 0.16 -0.01 -0.018 -0.0097 -0.013 -0.0024 0.0007 -0.0027 -0.011 0.0096 -0.005 0.048
event_type_Non-violent transfer of territory -0.082 -0.13 -0.006 0.13 -0.015 -0.029 -0.04 -0.061 0.0012 0.088 -0.046 0.0018 -0.031 -0.021 -0.015 -0.012 -0.081 -0.0064 -0.011 -0.018 -0.032 -0.008 -0.027 -0.022 -0.032 0.0032 0.42 -0.0064 -0.011 -0.0022 -0.0014 -0.0015 -0.03 0.024 0.14 -0.012 -0.0065 -0.0037 -0.0037 -0.05 -0.011 0.019 -0.02 -0.0016 -0.001 -0.023 -0.0035 -0.025 -0.012 -0.0027 -0.0092 -0.0049 -0.056 -0.0008 -0.0028 0.0027 -0.001 -0.0085 0.2 0.0067 -0.075 -0.014 -0.0086 1.0 -0.033 -0.079 -0.033 -0.077 0.031 -0.0042 -0.0022 -0.0099 -0.026 -0.0062 -0.0043 -0.0047 0.021 -0.0033 -0.031 -0.0021 0.0026 -0.022 -0.0045 -0.0042 -0.008 0.0068 -0.0047 -0.013 -0.022 -0.0033 -0.005 -0.0038 -0.0089 -0.0064 0.0066 -0.0057 -0.011 -0.0085 -0.0063 -0.0079 -0.025 -0.004 -0.0078 -0.01 0.33 0.026 -0.033 -0.017 -0.031 -0.0043 -0.0084 -0.0049 -0.021 -0.023 -0.011 -0.024
event_type_Remote violence -0.0061 0.06 -0.0019 -0.0044 0.064 -0.061 -0.083 -0.13 -0.00085 0.13 -0.023 0.044 0.088 -0.032 -0.032 -0.024 0.026 0.043 -0.024 0.01 0.091 -0.0072 -0.056 -0.045 0.051 0.043 -0.017 -0.014 0.012 -0.0053 -0.0029 -0.0031 -0.0094 0.095 0.093 -0.019 -0.0096 -0.0078 -0.0025 0.02 0.099 -0.0031 -0.036 -0.0014 -0.0021 -0.046 0.00087 -0.052 -0.024 -0.0055 -0.019 -0.01 -0.12 -0.006 -0.0047 -0.0031 -0.0021 0.059 -0.0097 -0.00073 -0.16 -0.03 -0.018 -0.033 1.0 -0.16 -0.069 -0.16 -0.022 -0.0087 -0.0046 -0.02 -0.042 -0.005 -0.039 -0.0068 -0.059 0.00087 0.043 -0.0043 0.018 -0.035 -0.0094 -0.0066 -0.017 -0.017 -0.0031 -0.022 -0.031 -0.0069 -0.023 0.14 -0.014 -0.013 0.04 -0.012 -0.02 -0.017 -0.012 -0.005 -0.025 -0.0014 -0.01 -0.012 -0.044 0.12 -0.064 -0.013 0.098 -0.00037 -0.0041 -0.01 -0.018 -0.037 -0.022 -0.048
event_type_Riots/Protests -0.019 0.21 -0.035 -0.25 -0.34 -0.17 0.47 0.75 -0.0098 -0.13 0.077 -0.24 -0.18 -0.12 0.19 0.036 -0.42 -0.05 -0.064 -0.24 -0.2 -0.061 0.34 0.16 -0.16 -0.05 -0.086 -0.036 -0.077 -0.037 0.017 -0.002 -0.16 -0.1 -0.083 -0.074 -0.044 0.047 0.0062 -0.26 -0.065 -0.023 -0.1 0.019 0.0031 -0.12 -0.017 0.32 0.15 0.022 -0.046 0.062 0.71 0.036 0.022 -0.0074 -0.0051 -0.046 -0.045 -0.013 -0.38 -0.072 -0.043 -0.079 -0.16 1.0 -0.17 -0.39 -0.07 0.037 0.019 0.081 -0.091 -0.0092 -0.049 -0.019 -0.1 -0.0025 0.13 -0.0017 -0.029 0.027 0.043 0.0082 0.043 0.025 0.00094 0.024 0.049 0.0002 0.0024 -0.058 0.044 0.04 -0.036 0.057 0.091 0.0023 0.064 0.004 0.03 -0.016 -0.026 0.048 -0.096 -0.2 0.33 -0.09 -0.11 0.023 0.022 0.044 0.19 -0.035 0.076 -0.037
event_type_Strategic development -0.022 0.02 -0.015 0.071 0.018 -0.038 -0.085 -0.13 0.054 0.016 -0.071 -0.059 0.011 -0.038 -0.032 -0.021 0.01 -0.0061 0.0025 -0.062 0.016 -0.014 -0.057 -0.042 0.11 0.0022 0.22 -0.0025 -0.021 -0.01 -0.0029 -0.0032 0.016 -0.028 0.25 -0.017 -0.014 -0.0079 -0.0053 -0.051 -0.0088 0.11 -0.034 -0.0034 -0.0022 -0.038 -0.0065 -0.053 -0.025 -0.0057 -0.019 -0.01 -0.12 -0.0062 -0.0059 0.042 0.031 -0.011 0.11 0.016 -0.16 -0.03 -0.018 -0.033 -0.069 -0.17 1.0 -0.16 -0.03 -0.0053 0.0034 -0.0046 0.14 0.0046 0.019 -0.0038 0.025 0.0081 0.0099 0.0058 -0.00093 -0.022 -0.00012 0.016 -0.0062 -0.005 0.0024 -0.011 -0.02 0.0048 -0.0076 -0.0032 -0.011 -0.0089 0.019 0.00056 0.0042 0.015 -0.013 -0.0021 -0.038 0.0099 0.0021 -0.0099 0.08 -0.035 -0.064 0.013 -0.01 0.011 -0.00043 -0.0068 -0.014 0.017 -0.0044 -0.0048
event_type_Violence against civilians 0.029 -0.064 0.0014 0.1 0.35 0.12 -0.15 -0.27 -0.0095 -0.079 -0.23 -0.23 -0.18 -0.12 -0.075 0.058 0.87 -0.077 -0.062 -0.23 -0.19 -0.059 -0.13 0.0065 0.25 -0.048 -0.083 -0.035 -0.074 -0.035 -0.0067 0.014 0.34 -0.096 -0.08 -0.072 -0.042 -0.018 0.019 0.56 -0.063 -0.022 -0.1 -0.0078 0.005 0.29 -0.017 -0.12 -0.056 -0.0015 0.12 -0.024 -0.28 -0.014 -0.00078 -0.0072 -0.005 0.072 -0.044 -0.013 -0.37 -0.07 -0.042 -0.077 -0.16 -0.39 -0.16 1.0 -0.03 -0.011 -0.0079 -0.03 0.071 0.0041 0.064 0.0041 0.024 0.0019 -0.085 0.0069 -0.0049 -0.008 -0.014 0.005 -0.008 2.7e-05 -0.012 -0.0013 0.019 0.0045 -0.013 -0.054 -0.0073 0.0048 -0.017 -0.021 -0.022 0.015 -0.018 -0.018 0.031 -0.0067 0.04 -0.017 -0.011 -0.034 -0.083 0.034 0.067 -0.0058 0.012 -0.011 -0.084 0.031 -0.0059 0.19
country_Angola -0.067 -0.25 0.075 0.0066 -0.063 -0.037 -0.043 -0.046 -0.0022 -0.024 -0.047 0.23 -0.037 -0.026 -0.016 -0.013 -0.04 -0.0057 -0.013 0.23 -0.039 -0.012 -0.028 -0.0088 -0.0084 0.0097 0.031 -0.008 -0.017 -0.0081 -0.0015 -0.0017 0.021 -0.02 -0.017 -0.015 -0.0084 -0.0042 -0.0041 -0.044 -0.014 -0.005 -0.023 -0.0018 -0.0011 -0.025 -0.0039 -0.027 -0.012 -0.003 -0.0097 -0.0055 -0.047 -0.0032 -0.0031 -0.0016 -0.0011 -0.0096 -0.0064 -0.0029 0.11 -0.012 -0.0083 0.031 -0.022 -0.07 -0.03 -0.03 1.0 -0.0047 -0.0025 -0.011 -0.03 -0.011 -0.022 -0.0093 -0.04 -0.0036 -0.034 -0.0023 -0.0069 -0.025 -0.005 -0.0047 -0.0089 -0.01 -0.0052 -0.015 -0.027 -0.0037 -0.013 -0.028 -0.011 -0.0072 -0.014 -0.0074 -0.013 -0.01 -0.0095 -0.0088 -0.04 -0.0075 -0.0087 -0.012 -0.024 -0.059 -0.037 -0.025 -0.039 -0.0048 -0.0094 -0.0055 -0.023 -0.025 -0.012 -0.027
country_Benin -0.016 0.01 -0.0018 -0.013 -0.012 -0.00049 -0.0034 0.043 -0.00052 -0.0068 -0.0044 -0.013 -0.0081 0.0021 -0.0027 0.0058 -0.014 -0.0043 0.00014 -0.013 -0.0086 -0.0032 0.0033 0.006 -0.0049 -0.0027 -0.0046 -0.0019 -0.0041 -0.002 -0.00037 -0.0004 -0.0085 -0.0053 -0.0044 -0.0039 -0.0023 -0.001 -0.00098 -0.0069 -0.0035 -0.0012 0.0046 -0.00043 -0.00027 -0.0031 -0.00093 -0.0059 -0.0031 -0.00071 -0.0025 -0.0013 0.044 -0.00078 -0.00075 -0.0004 -0.00027 -0.0025 -0.0024 -0.00071 -0.015 -0.0038 -0.0023 -0.0042 -0.0087 0.037 -0.0053 -0.011 -0.0047 1.0 -0.0006 -0.0027 -0.0073 -0.0025 -0.0053 -0.0022 -0.0095 -0.00087 -0.0082 -0.00056 -0.0017 -0.0059 -0.0012 -0.0011 -0.0021 -0.0025 -0.0013 -0.0035 -0.0064 -0.0009 -0.0031 -0.0068 -0.0027 -0.0017 -0.0035 -0.0018 -0.0031 -0.0025 -0.0023 -0.0021 -0.0095 -0.0018 -0.0021 -0.0028 -0.0057 -0.014 -0.0089 -0.006 -0.0094 -0.0012 -0.0023 -0.0013 -0.0056 -0.006 -0.003 -0.0065
country_Botswana -0.018 -0.0056 -0.001 -0.006 -0.0047 -0.0048 0.0093 0.015 -0.00028 -0.0019 0.002 -0.0067 -0.0041 -0.0035 0.0033 -0.0018 -0.0075 0.00039 -0.0018 -0.0069 -0.0044 -0.0017 0.0028 0.005 -0.0004 0.0029 -0.0024 -0.001 -0.0022 -0.001 -0.0002 -0.00022 -0.0045 -0.00064 0.00027 0.00082 -0.0012 -0.00053 -0.00052 -0.0043 -0.0018 -0.00064 -0.0029 -0.00023 -0.00015 -0.0033 -0.0005 0.0085 0.0057 -0.00038 -0.0013 -0.0007 0.014 -0.00041 -0.0004 -0.00021 -0.00015 -0.0013 -0.0013 -0.00038 -0.0092 -0.002 -0.0012 -0.0022 -0.0046 0.019 0.0034 -0.0079 -0.0025 -0.0006 1.0 -0.0014 -0.0039 -0.0013 -0.0028 -0.0012 -0.0051 -0.00046 -0.0044 -0.0003 -0.00089 -0.0032 -0.00064 -0.00061 -0.0011 -0.0013 -0.00067 -0.0019 -0.0034 -0.00048 -0.0016 -0.0036 -0.0014 -0.00092 -0.0018 -0.00095 -0.0016 -0.0013 -0.0012 -0.0011 -0.0051 -0.00096 -0.0011 -0.0015 -0.003 -0.0076 -0.0047 -0.0032 -0.005 -0.00062 -0.0012 -0.00071 -0.003 -0.0032 -0.0016 -0.0035
country_Burkina Faso 0.042 0.029 -0.0041 -0.025 -0.029 -0.0073 0.038 0.068 -0.0012 -0.015 0.0054 -0.028 -0.016 -0.0051 0.0028 -0.0055 -0.033 -0.0071 0.0064 -0.027 -0.017 -0.0036 0.015 0.0079 -0.0083 -0.0043 -0.0085 -0.0045 -0.0077 -0.0046 -0.00087 -0.00095 -0.016 -0.013 -0.0098 -0.004 -0.00099 0.0002 0.0003 -0.024 -0.0059 -0.0028 -0.0019 -0.001 -0.00064 -0.0076 -0.0022 0.033 0.0077 -0.0017 0.0099 -0.0031 0.067 -0.0018 -0.0018 -0.00093 -0.00064 -0.0047 -0.0035 -0.0017 -0.031 -0.009 -0.0054 -0.0099 -0.02 0.081 -0.0046 -0.03 -0.011 -0.0027 -0.0014 1.0 -0.017 -0.006 -0.012 -0.0053 -0.022 -0.0021 -0.019 -0.0013 -0.0039 -0.014 -0.0029 -0.0027 -0.005 -0.0058 -0.003 -0.0083 -0.015 -0.0021 -0.0072 -0.016 -0.0064 -0.0041 -0.0082 -0.0042 -0.0072 -0.0058 -0.0053 -0.005 -0.022 -0.0042 -0.0049 -0.0066 -0.013 -0.034 -0.021 -0.014 -0.022 -0.0027 -0.0053 -0.0031 -0.013 -0.014 -0.007 -0.015
country_Burundi -0.21 -0.02 -0.0039 0.056 0.058 -0.051 -0.04 -0.074 0.0023 -0.039 0.00018 -0.0073 -0.031 -0.041 -0.021 -0.011 0.12 -0.017 -0.0058 0.03 -0.016 -0.021 -0.024 -0.018 0.11 -0.012 -0.0098 0.009 -0.017 -0.012 -0.0024 -0.0026 0.058 -0.027 0.026 -0.011 -0.013 0.0082 0.0027 0.068 -0.02 -0.0077 -0.035 -0.0028 -0.0018 -0.03 -0.006 -0.025 -0.019 -0.0046 -0.011 -0.0085 -0.072 -0.005 -0.0048 -6.4e-05 -0.0018 -0.011 -0.014 -0.0017 -0.014 -0.02 -0.0088 -0.026 -0.042 -0.091 0.14 0.071 -0.03 -0.0073 -0.0039 -0.017 1.0 -0.016 -0.034 -0.014 -0.061 -0.0056 -0.053 -0.0036 -0.011 -0.038 -0.0078 -0.0073 -0.014 -0.016 -0.0081 -0.023 -0.041 -0.0058 -0.02 -0.044 -0.017 -0.011 -0.022 -0.011 -0.02 -0.016 -0.015 -0.014 -0.061 -0.012 -0.013 -0.018 -0.037 -0.092 -0.057 -0.039 -0.06 -0.0075 -0.015 -0.0085 -0.036 -0.039 -0.019 -0.042
country_Cameroon 0.04 0.036 0.00028 0.061 -0.014 -0.0066 -0.0052 -0.0065 -0.0012 -0.01 0.031 -0.017 -0.011 0.014 -0.0073 -0.0042 0.0035 -0.0026 -0.0077 0.017 -0.0018 -0.0048 0.0035 0.0024 -0.004 0.0074 0.0018 -0.0043 -0.0092 0.05 -0.00083 -0.00091 0.041 -0.011 -0.0061 -0.0061 0.003 -0.0023 -0.0022 -0.013 -0.007 -0.0027 0.0013 -0.00096 -0.00061 -0.011 -0.0021 -0.0086 -0.0052 -0.0016 -0.0033 -0.003 -0.0084 -0.0017 -0.0017 -0.00089 -0.00061 -0.0055 0.0025 -0.0016 0.012 -0.0072 -0.004 -0.0062 -0.005 -0.0092 0.0046 0.0041 -0.011 -0.0025 -0.0013 -0.006 -0.016 1.0 -0.012 -0.005 -0.021 -0.002 -0.018 -0.0013 -0.0037 -0.013 -0.0027 -0.0026 -0.0048 -0.0056 -0.0028 -0.0079 -0.014 -0.002 -0.0069 -0.015 -0.0061 -0.0039 -0.0078 -0.004 -0.0068 -0.0055 -0.0051 -0.0048 -0.021 -0.004 -0.0047 -0.0063 -0.013 -0.032 -0.02 -0.013 -0.021 -0.0026 -0.0051 -0.003 -0.013 -0.014 -0.0067 -0.015
country_Central African Republic 0.084 0.044 -0.003 0.0071 0.085 0.0021 -0.0088 -0.044 0.00011 0.023 -0.037 -0.041 0.029 0.00071 -0.0061 -0.014 0.055 0.068 0.0055 -0.042 -0.028 -0.011 -0.027 -0.024 -0.018 0.0042 -0.0057 0.00063 -0.011 0.0045 -0.0017 0.0047 0.027 -0.011 0.023 0.099 0.036 0.019 0.00081 0.049 0.066 0.0032 -0.013 0.0073 -0.0013 0.014 0.014 -0.0092 -0.0033 -0.0033 0.021 0.065 -0.04 -0.0036 0.021 0.0015 -0.0013 0.029 0.024 0.0099 -0.003 0.0083 -0.0072 -0.0043 -0.039 -0.049 0.019 0.064 -0.022 -0.0053 -0.0028 -0.012 -0.034 -0.012 1.0 -0.01 -0.044 -0.0041 -0.038 -0.0026 -0.0078 -0.028 -0.0056 -0.0053 -0.01 -0.012 -0.0059 -0.016 -0.03 -0.0042 -0.014 -0.032 -0.013 -0.008 -0.016 -0.0083 -0.014 -0.011 -0.011 -0.0099 -0.044 -0.0084 -0.0097 -0.013 -0.027 -0.066 -0.042 -0.028 -0.044 -0.0054 -0.011 -0.0062 -0.026 -0.028 -0.014 -0.03
country_Chad 0.036 -0.03 0.0042 0.0065 0.011 -0.0082 -0.013 -0.011 0.0048 0.0039 0.029 0.0088 -0.011 -0.0032 -0.0082 -0.0057 0.0027 0.0052 -0.00048 0.026 0.0077 -0.0045 -0.0058 0.0059 0.0036 0.026 -0.0043 0.0025 -0.0051 -0.00073 -0.00073 -0.0008 -0.006 -0.0082 -0.0066 -0.0054 -0.0033 -0.002 -0.0019 0.008 -0.0015 -0.0024 -0.0012 -0.00085 0.011 -0.0086 -0.0018 -0.011 -0.0062 -0.0014 -0.0024 -0.0026 -0.015 -0.0015 -0.0015 -0.00078 -0.00054 0.011 0.0003 0.0029 0.017 0.021 -0.0046 -0.0047 -0.0068 -0.019 -0.0038 0.0041 -0.0093 -0.0022 -0.0012 -0.0053 -0.014 -0.005 -0.01 1.0 -0.019 -0.0017 -0.016 -0.0011 -0.0033 -0.012 -0.0024 -0.0023 -0.0042 -0.0049 -0.0025 -0.007 -0.013 -0.0018 -0.0061 -0.013 -0.0054 -0.0034 -0.0069 -0.0035 -0.006 -0.0049 -0.0045 -0.0042 -0.019 -0.0036 -0.0041 -0.0055 -0.011 -0.028 -0.018 -0.012 -0.019 -0.0023 -0.0045 -0.0026 -0.011 -0.012 -0.0059 -0.013
country_Democratic Republic of Congo 0.58 -0.042 0.0075 0.11 0.014 -0.027 -0.045 -0.08 -0.0044 0.011 0.0099 0.027 0.05 -0.023 -0.025 -0.024 0.0098 0.028 0.007 0.028 0.047 -0.018 -0.024 -0.022 -0.021 -0.0039 0.046 0.049 0.033 -0.0077 0.0032 0.0023 0.084 0.01 0.013 0.012 0.023 -0.0069 -0.0067 -0.013 0.012 0.0015 -0.019 -0.0036 -0.0023 -0.027 0.021 -0.03 -0.021 -0.006 -0.013 -0.0024 -0.079 -0.0066 -0.0053 -0.0033 -0.0023 0.0021 0.017 0.037 0.049 0.064 0.017 0.021 -0.059 -0.1 0.025 0.024 -0.04 -0.0095 -0.0051 -0.022 -0.061 -0.021 -0.044 -0.019 1.0 -0.0074 -0.069 -0.0047 -0.014 -0.05 -0.01 -0.0096 -0.018 -0.021 -0.011 -0.03 -0.054 -0.0076 -0.026 -0.057 -0.023 -0.014 -0.029 -0.015 -0.026 -0.021 -0.019 -0.018 -0.08 -0.015 -0.018 -0.023 -0.048 -0.12 -0.075 -0.051 -0.079 -0.0098 -0.019 -0.011 -0.047 -0.051 -0.025 -0.055
country_Djibouti -0.026 -0.00087 -0.0012 -0.0054 -0.011 -0.007 -0.0044 -0.0008 -0.0004 -0.004 -0.0011 0.0065 -0.0053 -0.0038 0.0025 0.0045 0.0059 -0.0015 0.0043 0.0081 -0.0057 -8.3e-05 0.0035 0.0039 0.029 0.0009 -0.0035 -0.0015 -0.0032 -0.0015 -0.00029 -0.00031 -0.0026 -0.0041 -0.0034 -0.003 -0.0018 -0.00078 -0.00076 -0.0072 -0.0027 -0.00093 -0.0043 -0.00033 -0.00021 -0.0048 -0.00072 -0.0053 -0.0024 -0.00055 -0.0019 -0.001 -0.0024 0.0094 -0.00058 -0.00031 -0.00021 -0.0019 0.0014 -0.00055 -0.0025 -0.003 -0.0018 -0.0033 0.00087 -0.0025 0.0081 0.0019 -0.0036 -0.00087 -0.00046 -0.0021 -0.0056 -0.002 -0.0041 -0.0017 -0.0074 1.0 -0.0063 -0.00043 -0.0013 -0.0046 -0.00094 -0.00088 -0.0017 -0.0019 -0.00098 -0.0027 -0.005 -0.00069 -0.0024 -0.0053 -0.0021 -0.0013 -0.0027 -0.0014 -0.0024 -0.0019 -0.0018 -0.0016 -0.0073 -0.0014 -0.0016 -0.0022 -0.0044 -0.011 -0.0069 -0.0047 -0.0073 -0.0009 -0.0017 -0.001 -0.0043 -0.0047 -0.0023 -0.005
country_Egypt -0.12 0.11 -0.0096 -0.081 -0.049 -0.055 0.07 0.094 -0.0021 -0.047 0.029 -0.067 0.083 -0.038 0.086 0.04 -0.093 -0.017 -0.023 -0.066 0.11 -0.018 0.067 0.056 -0.0089 -0.018 -0.029 -0.014 -0.029 -0.011 -0.0027 -0.0029 -0.051 -0.038 0.033 -0.022 -0.015 0.0033 0.019 -0.062 -0.018 -0.0087 -0.033 0.0072 0.0012 -0.04 -0.0048 -0.015 0.097 0.05 0.014 0.0071 0.078 0.0079 0.016 -0.00064 -0.002 -0.016 -0.017 -0.0051 -0.054 -0.027 -0.017 -0.031 0.043 0.13 0.0099 -0.085 -0.034 -0.0082 -0.0044 -0.019 -0.053 -0.018 -0.038 -0.016 -0.069 -0.0063 1.0 -0.0041 -0.012 -0.043 -0.0088 -0.0083 -0.016 -0.018 -0.0091 -0.026 -0.046 -0.0065 -0.022 -0.049 -0.02 -0.012 -0.025 -0.013 -0.022 -0.018 -0.016 -0.015 -0.069 -0.013 -0.015 -0.02 -0.041 -0.1 -0.065 -0.044 -0.068 -0.0084 -0.016 -0.0096 -0.041 -0.044 -0.022 -0.047
country_Equatorial Guinea -0.025 -0.0072 -0.00081 -0.0054 -0.0037 -0.0045 -0.0041 0.0011 -0.00026 -0.0034 -0.0019 -0.0018 -0.0037 -0.0032 -0.0021 0.0094 0.0062 -0.0021 0.0019 -0.0021 -0.0014 -0.0016 -0.0018 0.0058 0.021 -0.0013 -0.0023 -0.00096 -0.002 -0.00097 -0.00018 -0.0002 -0.0027 -0.0026 -0.0022 -0.002 -0.0011 -0.0005 -0.00049 -0.0026 -0.0017 -0.0006 -0.0027 -0.00021 -0.00014 -0.003 -0.00046 -0.0034 -0.0015 -0.00035 -0.0012 -0.00066 0.0012 -0.00039 -0.00037 -0.0002 -0.00014 -0.0012 -0.0012 -0.00035 -0.0043 -0.0019 -0.0011 -0.0021 -0.0043 -0.0017 0.0058 0.0069 -0.0023 -0.00056 -0.0003 -0.0013 -0.0036 -0.0013 -0.0026 -0.0011 -0.0047 -0.00043 -0.0041 1.0 -0.00083 -0.0029 -0.0006 -0.00056 -0.0011 -0.0012 -0.00062 -0.0017 -0.0032 -0.00044 -0.0015 -0.0034 -0.0013 -0.00085 -0.0017 -0.00089 -0.0015 -0.0012 -0.0011 -0.0011 -0.0047 -0.00089 -0.001 -0.0014 -0.0028 -0.0071 -0.0044 -0.003 -0.0047 -0.00058 -0.0011 -0.00066 -0.0028 -0.003 -0.0015 -0.0032
country_Eritrea -0.049 -0.076 0.12 -0.014 -0.019 -0.013 -0.016 -0.021 -0.00077 0.16 0.06 -0.016 -0.013 -0.009 -0.0051 -0.0049 -0.00077 -0.0015 -0.0014 -0.015 -0.011 -0.0035 -0.011 -0.0086 0.021 0.34 -0.0031 -0.0029 -0.006 -0.0029 -0.00055 -0.0006 -0.011 -0.0039 -0.0037 -0.0058 -0.0034 -0.0015 -0.0015 -0.015 -0.0051 -0.0018 -0.0081 -0.00063 -0.0004 -0.0091 -0.0014 -0.01 -0.0033 -0.0011 -0.0036 -0.002 -0.019 -0.0012 -0.0011 -0.00058 -0.0004 0.057 0.015 -0.001 0.025 0.0019 0.00018 0.0026 0.018 -0.029 -0.00093 -0.0049 -0.0069 -0.0017 -0.00089 -0.0039 -0.011 -0.0037 -0.0078 -0.0033 -0.014 -0.0013 -0.012 -0.00083 1.0 -0.0088 -0.0018 -0.0017 -0.0032 -0.0037 -0.0019 -0.0052 -0.0095 -0.0013 -0.0045 -0.01 -0.004 -0.0025 -0.0051 -0.0026 -0.0045 -0.0036 -0.0034 -0.0031 -0.014 -0.0027 -0.0031 -0.0041 -0.0085 -0.021 -0.013 -0.0089 -0.014 -0.0017 -0.0033 -0.002 -0.0083 -0.0089 -0.0044 -0.0096
country_Ethiopia 0.095 0.039 0.014 -0.024 -0.079 -0.015 -0.029 0.057 -0.00049 -0.03 0.0028 0.036 -0.043 0.0097 0.0018 0.16 -0.056 -0.0031 -0.0041 0.069 -0.045 0.022 -0.012 0.12 0.038 0.012 -0.017 -0.01 -0.019 -0.0097 -0.0019 -0.0021 -0.038 -0.025 -0.018 -0.019 -0.0091 -0.0053 0.002 -0.059 -0.017 -0.0053 0.0047 -0.0023 -0.0014 -0.023 -0.0011 -0.012 -0.013 -0.0038 -0.0091 -0.007 0.047 -0.0041 -0.0039 -0.0021 -0.0014 -0.0043 -0.012 -0.002 0.029 -0.011 -0.011 -0.022 -0.035 0.027 -0.022 -0.008 -0.025 -0.0059 -0.0032 -0.014 -0.038 -0.013 -0.028 -0.012 -0.05 -0.0046 -0.043 -0.0029 -0.0088 1.0 -0.0064 -0.006 -0.011 -0.013 -0.0066 -0.019 -0.034 -0.0047 -0.016 -0.036 -0.014 -0.0091 -0.018 -0.0094 -0.016 -0.013 -0.012 -0.011 -0.05 -0.0095 -0.011 -0.015 -0.03 -0.075 -0.047 -0.032 -0.049 -0.0061 -0.012 -0.007 -0.03 -0.032 -0.016 -0.034
country_Gabon -0.035 0.015 -0.002 -0.013 -0.016 -0.0091 0.017 0.037 -0.00056 -0.0064 0.013 -0.013 -0.01 -0.007 -0.00032 -0.00019 -0.016 -0.0046 -0.0037 -0.014 -0.0095 -0.0035 0.019 0.022 0.0026 -0.0028 -0.0024 -0.0021 -0.0044 -0.0021 -0.0004 -0.00043 -0.0091 -0.0046 -0.0021 -0.0042 -0.0025 -0.0011 0.0047 -0.013 -0.0037 -0.0013 -0.0059 -0.00046 -0.00029 -0.0056 -0.001 0.01 -0.0033 -0.00077 -0.0026 -0.0014 0.028 0.0064 -0.0008 -0.00042 -0.00029 -0.0026 -0.0026 -0.00076 -0.021 -0.0041 -0.0025 -0.0045 -0.0094 0.043 -0.00012 -0.014 -0.005 -0.0012 -0.00064 -0.0029 -0.0078 -0.0027 -0.0056 -0.0024 -0.01 -0.00094 -0.0088 -0.0006 -0.0018 -0.0064 1.0 -0.0012 -0.0023 -0.0027 -0.0014 -0.0038 -0.0069 -0.00096 -0.0033 -0.0073 -0.0029 -0.0018 -0.0037 -0.0019 -0.0033 -0.0026 -0.0024 -0.0023 -0.01 -0.0019 -0.0022 -0.003 -0.0061 -0.015 -0.0096 -0.0064 -0.01 -0.0012 -0.0024 -0.0014 -0.006 -0.0065 -0.0032 -0.007
country_Gambia -0.068 -0.0031 -0.0017 -0.011 -0.0032 -0.0084 0.005 0.0081 -0.00053 -0.0031 -0.0021 -0.012 -0.0069 -0.0056 0.0075 0.0039 0.0073 0.007 0.012 -0.012 -0.0087 -0.0033 0.0014 0.0038 0.02 0.0019 -0.0033 -0.0019 -0.0041 -0.002 -0.00037 -0.00041 -0.0078 -0.003 -0.0044 0.0022 -0.0023 -0.001 -0.00099 0.00059 1.1e-05 -0.0012 -0.0044 -0.00043 -0.00028 -0.0062 -0.00094 -0.0041 0.0066 -0.00072 0.017 0.012 0.0081 -0.00078 -0.00075 -0.0004 -0.00028 -2.9e-05 0.0025 -0.00071 -0.015 -0.0039 -0.0023 -0.0042 -0.0066 0.0082 0.016 0.005 -0.0047 -0.0011 -0.00061 -0.0027 -0.0073 -0.0026 -0.0053 -0.0023 -0.0096 -0.00088 -0.0083 -0.00056 -0.0017 -0.006 -0.0012 1.0 -0.0022 -0.0025 -0.0013 -0.0036 -0.0065 -0.0009 -0.0031 -0.0069 -0.0027 -0.0017 -0.0035 -0.0018 -0.0031 -0.0025 -0.0023 -0.0021 -0.0096 -0.0018 -0.0021 -0.0028 -0.0058 -0.014 -0.009 -0.0061 -0.0095 -0.0012 -0.0023 -0.0013 -0.0057 -0.0061 -0.003 -0.0066
country_Ghana -0.063 0.027 -0.0031 -0.024 -0.014 0.03 0.021 0.036 -0.00099 -0.012 -0.0031 -0.024 -0.013 0.025 0.017 -0.0035 -0.012 -0.0074 -0.0056 -0.024 -0.015 -0.0042 0.016 -0.00075 -0.0077 -0.005 -0.0087 -0.0037 -0.0078 -0.0037 -0.0007 -0.00077 -0.015 -0.01 -0.0054 0.0016 0.0025 -0.0019 -0.0019 -0.0087 -0.0056 0.00038 0.032 -0.00081 -0.00052 0.014 0.0017 0.0008 0.023 -0.0014 0.0096 -0.0025 0.038 -0.0015 -0.0014 -0.00075 -0.00052 -0.0034 -0.0046 -0.0013 -0.017 -0.0073 -0.003 -0.008 -0.017 0.043 -0.0062 -0.008 -0.0089 -0.0021 -0.0011 -0.005 -0.014 -0.0048 -0.01 -0.0042 -0.018 -0.0017 -0.016 -0.0011 -0.0032 -0.011 -0.0023 -0.0022 1.0 -0.0047 -0.0024 -0.0067 -0.012 -0.0017 -0.0058 -0.013 -0.0051 -0.0033 -0.0066 -0.0034 -0.0058 -0.0047 -0.0043 -0.004 -0.018 -0.0034 -0.0039 -0.0053 -0.011 -0.027 -0.017 -0.011 -0.018 -0.0022 -0.0043 -0.0025 -0.011 -0.011 -0.0056 -0.012
country_Guinea 0.04 -0.042 -0.0009 0.0033 -0.018 -0.01 0.041 -0.0055 -0.0011 -0.011 0.021 -0.016 -0.008 -0.0042 0.037 0.011 -0.0044 -0.0055 0.012 -0.013 -0.0036 -0.0062 0.042 0.011 0.01 -0.0048 0.0072 -0.0042 -0.0055 -0.0043 -0.00081 -0.00089 0.0082 -0.0074 -0.0033 -0.0044 -0.0051 -0.0022 -0.0022 -0.016 -0.006 -0.0026 -0.0007 -0.00094 -0.0006 -0.0098 -0.002 0.0021 0.036 -0.0016 0.014 -0.0029 -0.0072 -0.0017 -0.0016 -0.00087 -0.0006 -0.0031 -0.0018 -0.0015 -0.014 0.00036 -0.005 0.0068 -0.017 0.025 -0.005 2.7e-05 -0.01 -0.0025 -0.0013 -0.0058 -0.016 -0.0056 -0.012 -0.0049 -0.021 -0.0019 -0.018 -0.0012 -0.0037 -0.013 -0.0027 -0.0025 -0.0047 1.0 -0.0028 -0.0077 -0.014 -0.002 -0.0067 -0.015 -0.0059 -0.0038 -0.0076 -0.0039 -0.0067 -0.0054 -0.005 -0.0047 -0.021 -0.0039 -0.0046 -0.0061 -0.013 -0.031 -0.02 -0.013 -0.021 -0.0025 -0.0049 -0.0029 -0.012 -0.013 -0.0065 -0.014
country_Guinea-Bissau -0.056 -0.037 -0.00015 -0.0075 -0.013 -0.0095 -0.0049 0.006 -0.00058 -0.00092 0.031 -0.0082 -0.0083 -0.0064 -0.0046 -0.0021 -0.013 0.0003 0.12 -0.0057 -0.006 -0.0036 -0.0033 -0.00068 0.001 0.0052 -0.0015 0.00065 -0.0046 -0.0022 -0.00041 -0.00045 -0.0081 -0.0038 -0.0024 -0.0044 -0.0026 -0.0011 -0.0011 -0.0097 -0.0039 -0.0013 -0.0051 -0.00048 -0.00031 -0.0069 -0.001 -0.0034 -0.0035 -0.0008 -0.00055 0.0026 0.0059 -0.00087 0.0064 -0.00044 -0.00031 0.0017 0.0018 -0.00079 0.012 0.0071 0.0022 -0.0047 -0.0031 0.00094 0.0024 -0.012 -0.0052 -0.0013 -0.00067 -0.003 -0.0081 -0.0028 -0.0059 -0.0025 -0.011 -0.00098 -0.0091 -0.00062 -0.0019 -0.0066 -0.0014 -0.0013 -0.0024 -0.0028 1.0 -0.0039 -0.0071 -0.001 -0.0034 -0.0076 -0.003 -0.0019 -0.0039 -0.002 -0.0034 -0.0028 -0.0025 -0.0024 -0.011 -0.002 -0.0023 -0.0031 -0.0064 -0.016 -0.0099 -0.0067 -0.01 -0.0013 -0.0025 -0.0015 -0.0063 -0.0067 -0.0033 -0.0073
country_Ivory Coast 0.0032 -0.042 -0.0029 -0.031 -0.011 -0.0043 0.021 0.008 0.0021 -0.0067 -0.0079 -0.017 0.015 0.0048 0.011 0.011 -0.0079 0.0046 0.011 -0.026 0.0096 -0.0034 0.0099 0.007 0.024 0.0043 -0.014 0.0092 -0.0026 -0.0061 0.0041 -0.0013 -0.019 -0.01 -0.0079 0.0033 -0.0047 -0.0031 0.0009 -0.0093 0.0086 -0.0038 0.0083 -0.0013 -0.00085 -0.009 -0.0029 0.022 0.0042 0.006 0.0019 0.0033 0.0067 0.01 -0.0023 0.0037 -0.00085 -0.0013 -0.0043 0.00057 -0.0033 0.022 -0.0072 -0.013 -0.022 0.024 -0.011 -0.0013 -0.015 -0.0035 -0.0019 -0.0083 -0.023 -0.0079 -0.016 -0.007 -0.03 -0.0027 -0.026 -0.0017 -0.0052 -0.019 -0.0038 -0.0036 -0.0067 -0.0077 -0.0039 1.0 -0.02 -0.0028 -0.0096 -0.021 -0.0084 -0.0054 -0.011 -0.0056 -0.0095 -0.0077 -0.0071 -0.0066 -0.03 -0.0056 -0.0065 -0.0087 -0.018 -0.044 -0.028 -0.019 -0.029 -0.0036 -0.0071 -0.0041 -0.018 -0.019 -0.0093 -0.02
country_Kenya 0.0058 -0.06 -0.0068 -0.049 -0.016 0.12 0.054 0.014 -0.003 -0.028 0.00093 -0.049 -0.018 0.051 0.02 -0.007 0.0091 -0.02 -0.016 -0.048 -0.018 0.044 0.035 0.0033 -0.026 -0.0092 -0.015 -0.0069 -0.022 -0.0099 0.0039 -0.0023 -0.027 -0.027 -0.013 0.008 -0.005 0.0086 2.5e-05 -0.0061 -0.018 0.012 0.051 -0.0024 0.0025 0.1 0.0089 0.047 0.0042 -0.0025 0.012 -0.0033 0.011 0.0012 -0.0042 -0.0022 -0.0016 -0.0045 -0.013 -0.004 -0.022 -0.021 -0.011 -0.022 -0.031 0.049 -0.02 0.019 -0.027 -0.0064 -0.0034 -0.015 -0.041 -0.014 -0.03 -0.013 -0.054 -0.005 -0.046 -0.0032 -0.0095 -0.034 -0.0069 -0.0065 -0.012 -0.014 -0.0071 -0.02 1.0 -0.0051 -0.017 -0.039 -0.015 -0.0098 -0.02 -0.01 -0.017 -0.014 -0.013 -0.012 -0.054 -0.01 -0.012 -0.016 -0.032 -0.081 -0.051 -0.034 -0.053 -0.0066 -0.013 -0.0075 -0.032 -0.034 -0.017 -0.037
country_Lesotho -0.053 -0.022 -0.00099 -0.0098 0.0096 -0.0054 0.0015 -0.002 -0.00041 0.004 0.0092 -0.01 -0.0072 -0.004 -0.0014 0.0019 0.0039 0.0056 0.022 -0.01 -0.00038 -0.0026 -0.0013 -0.00054 0.0025 0.027 -0.0036 -0.0015 -0.0032 -0.0016 -0.00029 -0.00032 -0.0067 -0.0027 -0.0017 -0.0012 -0.0018 -0.0008 -0.00078 0.0099 -0.0027 -0.00095 -0.003 -0.00034 -0.00022 -0.0036 -0.00074 0.0039 -3.2e-06 -0.00057 -0.002 -0.0011 -0.0011 -0.00062 -0.00059 -0.00031 -0.00022 0.0012 0.0012 -0.00056 -0.0012 -0.001 -0.0018 -0.0033 -0.0069 0.0002 0.0048 0.0045 -0.0037 -0.0009 -0.00048 -0.0021 -0.0058 -0.002 -0.0042 -0.0018 -0.0076 -0.00069 -0.0065 -0.00044 -0.0013 -0.0047 -0.00096 -0.0009 -0.0017 -0.002 -0.001 -0.0028 -0.0051 1.0 -0.0024 -0.0054 -0.0021 -0.0014 -0.0028 -0.0014 -0.0024 -0.002 -0.0018 -0.0017 -0.0075 -0.0014 -0.0017 -0.0022 -0.0045 -0.011 -0.0071 -0.0048 -0.0075 -0.00092 -0.0018 -0.0011 -0.0045 -0.0048 -0.0024 -0.0052
country_Liberia -0.09 -0.068 -0.0042 0.049 -0.037 -0.011 0.0036 0.00068 -0.0014 -0.0029 0.037 0.017 -0.021 -0.01 -0.0036 -0.0017 -0.02 0.014 -0.0073 0.066 -0.026 -0.0053 0.0029 0.00023 0.0023 0.012 -0.007 0.0052 -0.0061 0.026 -0.001 -0.0011 0.0012 -0.0036 -0.0078 -0.0072 -0.0044 -0.0027 -0.0027 -0.027 -0.0055 -0.0014 -0.012 -0.0012 -0.00074 -0.0084 -0.0025 0.0011 -0.0041 -0.0019 0.007 0.0049 -0.00085 -0.0021 0.028 -0.0011 -0.00074 0.008 -0.0047 0.0076 -0.00037 0.074 -0.0013 -0.005 -0.023 0.0024 -0.0076 -0.013 -0.013 -0.0031 -0.0016 -0.0072 -0.02 -0.0069 -0.014 -0.0061 -0.026 -0.0024 -0.022 -0.0015 -0.0045 -0.016 -0.0033 -0.0031 -0.0058 -0.0067 -0.0034 -0.0096 -0.017 -0.0024 1.0 -0.018 -0.0073 -0.0047 -0.0094 -0.0048 -0.0083 -0.0067 -0.0062 -0.0058 -0.026 -0.0049 -0.0056 -0.0076 -0.016 -0.039 -0.024 -0.016 -0.026 -0.0032 -0.0061 -0.0036 -0.015 -0.016 -0.0081 -0.018
country_Libya 0.11 0.12 -0.004 -0.027 0.029 0.02 -0.043 -0.032 0.023 0.0057 -0.031 -0.0072 0.13 0.063 -0.021 -0.013 -0.05 -0.007 -0.02 -0.015 0.091 0.022 -0.037 -0.03 -0.033 0.06 -0.013 -0.012 0.042 0.039 -0.0022 0.0027 -0.034 -0.017 0.056 0.029 0.055 0.0033 0.018 -0.0061 -0.0069 0.029 0.037 0.0096 0.006 -0.024 -3.5e-06 -0.02 -0.016 -0.0028 -0.0085 -0.0064 -0.025 0.00066 -0.0017 0.027 0.006 -0.0016 -0.0085 -0.0028 0.018 0.04 -0.0097 -0.0038 0.14 -0.058 -0.0032 -0.054 -0.028 -0.0068 -0.0036 -0.016 -0.044 -0.015 -0.032 -0.013 -0.057 -0.0053 -0.049 -0.0034 -0.01 -0.036 -0.0073 -0.0069 -0.013 -0.015 -0.0076 -0.021 -0.039 -0.0054 -0.018 1.0 -0.016 -0.01 -0.021 -0.011 -0.018 -0.015 -0.014 -0.013 -0.057 -0.011 -0.013 -0.017 -0.034 -0.086 -0.054 -0.036 -0.057 -0.007 -0.014 -0.008 -0.034 -0.036 -0.018 -0.039
country_Madagascar -0.2 0.013 -0.0029 -0.031 0.012 -0.0022 0.028 0.029 -0.0012 -0.016 0.026 -0.03 -0.0079 -0.00029 0.00062 -0.0011 -0.01 -0.0085 -0.0037 -0.031 0.014 -0.0062 0.032 0.0087 -0.0031 -0.0064 -0.011 -0.0046 -0.0098 -0.0047 -0.00088 -0.00097 -0.02 -0.013 -0.0023 -0.0081 0.037 0.013 -0.0024 -8.5e-05 -0.0068 -0.0029 -0.012 0.0049 -0.00065 0.0087 -0.0022 0.004 0.0033 -0.0017 -0.0018 -0.0012 0.027 -0.0019 -0.0018 -0.00095 -0.00065 -0.0059 -0.0058 -0.0017 -0.015 -0.0092 -0.0055 -0.0089 -0.014 0.044 -0.011 -0.0073 -0.011 -0.0027 -0.0014 -0.0064 -0.017 -0.0061 -0.013 -0.0054 -0.023 -0.0021 -0.02 -0.0013 -0.004 -0.014 -0.0029 -0.0027 -0.0051 -0.0059 -0.003 -0.0084 -0.015 -0.0021 -0.0073 -0.016 1.0 -0.0041 -0.0083 -0.0043 -0.0073 -0.0059 -0.0054 -0.0051 -0.023 -0.0043 -0.005 -0.0067 -0.014 -0.034 -0.021 -0.014 -0.022 -0.0028 -0.0054 -0.0032 -0.013 -0.014 -0.0071 -0.016
country_Malawi -0.05 0.0089 -0.0028 -0.02 -0.01 0.0092 0.034 0.016 -0.0008 -0.01 0.0052 -0.019 -0.011 -0.01 0.024 0.00088 0.0021 -0.0056 -0.0052 -0.02 -0.014 -0.0025 0.033 0.0088 0.0035 -0.004 -0.007 -0.0029 -0.0062 -0.003 -0.00056 -0.00062 -0.013 -0.0081 -0.0067 -0.0009 -0.0018 0.0024 0.0025 -0.0032 -0.0053 -0.0018 -0.0084 -0.00065 -0.00042 0.021 -0.0014 0.022 0.0056 -0.0011 0.011 0.00098 0.012 0.0039 -0.0011 -0.0006 -0.00042 -0.0037 -0.0037 -0.0011 -0.028 -0.0059 -0.0035 -0.0064 -0.013 0.04 -0.0089 0.0048 -0.0072 -0.0017 -0.00092 -0.0041 -0.011 -0.0039 -0.008 -0.0034 -0.014 -0.0013 -0.012 -0.00085 -0.0025 -0.0091 -0.0018 -0.0017 -0.0033 -0.0038 -0.0019 -0.0054 -0.0098 -0.0014 -0.0047 -0.01 -0.0041 1.0 -0.0053 -0.0027 -0.0047 -0.0038 -0.0035 -0.0032 -0.014 -0.0027 -0.0032 -0.0042 -0.0087 -0.022 -0.014 -0.0092 -0.014 -0.0018 -0.0034 -0.002 -0.0085 -0.0092 -0.0045 -0.0099
country_Mali 0.11 0.051 -0.0028 0.061 0.0086 -0.011 -0.022 -0.024 -0.0016 0.037 0.011 0.0054 0.00035 -0.004 -0.01 -0.0091 -0.015 0.075 -0.0012 0.001 0.0017 -0.0075 -0.017 -0.013 -0.011 -0.0082 0.019 0.032 0.023 0.0042 0.026 0.023 0.014 0.058 0.0043 -0.0034 0.0006 -0.0031 0.001 -0.015 0.056 -0.0004 -0.0036 -0.0013 -0.00084 -0.011 0.0014 -0.014 -0.0089 -0.0022 -0.0036 0.0064 -0.022 -0.0024 -0.0023 -0.0012 -0.00084 0.0022 0.025 0.0091 0.015 0.017 -0.0019 0.0066 0.04 -0.036 0.019 -0.017 -0.014 -0.0035 -0.0018 -0.0082 -0.022 -0.0078 -0.016 -0.0069 -0.029 -0.0027 -0.025 -0.0017 -0.0051 -0.018 -0.0037 -0.0035 -0.0066 -0.0076 -0.0039 -0.011 -0.02 -0.0028 -0.0094 -0.021 -0.0083 -0.0053 1.0 -0.0055 -0.0094 -0.0076 -0.007 -0.0065 -0.029 -0.0055 -0.0064 -0.0086 -0.018 -0.044 -0.027 -0.018 -0.029 -0.0036 -0.0069 -0.0041 -0.017 -0.019 -0.0091 -0.02
country_Mauritania -0.025 0.011 -0.0029 -0.014 -0.026 -0.014 0.0063 0.064 -0.00083 -0.0084 0.03 -0.014 -0.014 -0.0097 -0.0028 -0.0018 -0.025 -0.0032 0.0025 -0.011 -0.014 -0.0051 0.017 0.055 -0.0019 -0.0013 -0.0064 -0.0031 -0.0065 -0.0031 -0.00058 -0.00064 -0.011 -0.0062 -0.0061 -0.0062 -0.0037 -0.0016 -0.0016 -0.02 -0.0055 -0.0019 -0.008 -0.00068 -0.00043 -0.0097 -0.0015 -0.0084 -0.0012 -0.0011 0.00077 -0.0021 0.042 -0.0012 -0.0012 -0.00063 -0.00043 -0.00076 -0.0022 -0.0011 -0.026 -0.0041 -0.0036 -0.0057 -0.012 0.057 0.00056 -0.021 -0.0074 -0.0018 -0.00095 -0.0042 -0.011 -0.004 -0.0083 -0.0035 -0.015 -0.0014 -0.013 -0.00089 -0.0026 -0.0094 -0.0019 -0.0018 -0.0034 -0.0039 -0.002 -0.0056 -0.01 -0.0014 -0.0048 -0.011 -0.0043 -0.0027 -0.0055 1.0 -0.0048 -0.0039 -0.0036 -0.0034 -0.015 -0.0028 -0.0033 -0.0044 -0.009 -0.023 -0.014 -0.0095 -0.015 -0.0018 -0.0036 -0.0021 -0.0089 -0.0095 -0.0047 -0.01
country_Morocco -0.14 0.022 -0.0051 -0.033 -0.041 -0.024 0.01 0.11 -0.0014 -0.014 0.017 -0.032 -0.02 -0.018 0.0053 0.047 -0.043 -0.0079 -0.0093 -0.032 -0.02 -0.0088 0.023 0.077 -0.0067 -0.0063 -0.0094 -0.0052 -0.01 -0.0053 -0.001 -0.0011 -0.022 -0.014 -0.011 -0.0089 -0.0063 -0.0027 -0.00037 -0.032 -0.0087 -0.0032 -0.015 -0.0012 -0.00074 -0.016 -0.0025 -0.011 -0.00036 0.011 0.0098 0.0049 0.087 0.027 0.00099 -0.0011 -0.00074 -0.002 -0.0056 0.0045 -0.052 -0.01 -0.0062 -0.011 -0.02 0.091 0.0042 -0.022 -0.013 -0.0031 -0.0016 -0.0072 -0.02 -0.0068 -0.014 -0.006 -0.026 -0.0024 -0.022 -0.0015 -0.0045 -0.016 -0.0033 -0.0031 -0.0058 -0.0067 -0.0034 -0.0095 -0.017 -0.0024 -0.0083 -0.018 -0.0073 -0.0047 -0.0094 -0.0048 1.0 -0.0067 -0.0061 -0.0057 -0.026 -0.0049 -0.0056 -0.0075 -0.015 -0.038 -0.024 -0.016 -0.025 -0.0031 -0.0061 -0.0036 -0.015 -0.016 -0.008 -0.018
country_Mozambique 0.039 0.016 -0.0036 -0.029 0.054 -0.019 0.018 -0.0084 -0.0011 -0.012 0.022 -0.027 0.0021 -0.014 0.0025 -0.0023 0.015 -0.0068 -0.0075 -0.028 0.038 -0.0062 0.0093 -0.0042 0.0021 -0.0037 -0.01 -0.0042 -0.0089 -0.0043 -0.00081 -0.00088 -0.018 -0.012 0.013 -0.0029 -0.0051 0.037 0.00068 0.035 -0.0051 -0.0026 -0.012 -0.00094 -0.0006 -0.012 -0.002 0.0026 0.0058 -0.0016 0.013 -0.00079 -0.0072 -0.0017 -0.0016 -0.00086 -0.0006 -0.002 -0.003 -0.0015 -0.011 -0.0062 -0.0026 -0.0085 -0.017 0.0023 0.015 0.015 -0.01 -0.0025 -0.0013 -0.0058 -0.016 -0.0055 -0.011 -0.0049 -0.021 -0.0019 -0.018 -0.0012 -0.0036 -0.013 -0.0026 -0.0025 -0.0047 -0.0054 -0.0028 -0.0077 -0.014 -0.002 -0.0067 -0.015 -0.0059 -0.0038 -0.0076 -0.0039 -0.0067 1.0 -0.005 -0.0046 -0.021 -0.0039 -0.0045 -0.0061 -0.012 -0.031 -0.019 -0.013 -0.021 -0.0025 -0.0049 -0.0029 -0.012 -0.013 -0.0065 -0.014
country_Namibia -0.032 -0.01 -0.0036 -0.0097 -0.028 -0.014 -0.0014 0.077 -0.001 -0.0034 -0.017 -0.01 -0.016 -0.011 -0.0039 -0.004 -0.02 -0.0065 -0.0069 -0.015 -0.02 -0.0065 -0.0014 -0.0036 -0.0081 -0.0042 -0.0045 0.016 -0.0082 -0.0039 -0.00074 -0.00081 -0.0019 -0.0078 -0.0061 -0.004 -0.0047 0.01 -0.002 -0.02 -0.0061 -0.0024 -0.0088 -0.00086 -0.00055 -0.0088 -0.0019 -0.0049 -0.0033 -0.0014 0.0098 -0.00038 0.081 -0.0016 -0.0015 -0.0008 -0.00055 0.015 -0.0036 -0.0014 -0.027 -0.0069 -0.0046 -0.0063 -0.012 0.064 -0.013 -0.018 -0.0095 -0.0023 -0.0012 -0.0053 -0.015 -0.0051 -0.011 -0.0045 -0.019 -0.0018 -0.016 -0.0011 -0.0034 -0.012 -0.0024 -0.0023 -0.0043 -0.005 -0.0025 -0.0071 -0.013 -0.0018 -0.0062 -0.014 -0.0054 -0.0035 -0.007 -0.0036 -0.0061 -0.005 1.0 -0.0043 -0.019 -0.0036 -0.0042 -0.0056 -0.012 -0.029 -0.018 -0.012 -0.019 -0.0023 -0.0045 -0.0027 -0.011 -0.012 -0.006 -0.013
country_Niger 0.066 -0.0083 -0.0004 0.035 -0.022 -0.011 0.0043 0.0015 -0.00098 -0.0063 0.039 -0.0094 0.0079 -0.0031 -0.0047 -0.0053 -0.019 -0.0066 0.051 0.016 -0.0059 -0.0021 0.015 -0.0052 -0.0099 -0.0013 -0.0043 -0.0036 0.029 0.0046 -0.00069 -0.00076 0.0079 -0.0075 -0.0061 -0.0049 -0.003 -0.0019 -0.0019 -0.018 -0.0037 -0.0023 -0.0037 -0.00081 -0.00051 -0.01 -0.0018 -0.0059 -0.0058 -0.0013 0.00061 -0.0025 0.003 -0.0015 -0.0014 -0.00074 -0.00051 -0.0046 0.0021 -0.0013 0.012 0.037 -0.0043 -0.0079 -0.005 0.004 -0.0021 -0.018 -0.0088 -0.0021 -0.0011 -0.005 -0.014 -0.0048 -0.0099 -0.0042 -0.018 -0.0016 -0.015 -0.0011 -0.0031 -0.011 -0.0023 -0.0021 -0.004 -0.0047 -0.0024 -0.0066 -0.012 -0.0017 -0.0058 -0.013 -0.0051 -0.0032 -0.0065 -0.0034 -0.0057 -0.0046 -0.0043 1.0 -0.018 -0.0034 -0.0039 -0.0052 -0.011 -0.027 -0.017 -0.011 -0.018 -0.0022 -0.0042 -0.0025 -0.011 -0.011 -0.0056 -0.012
country_Nigeria -0.13 0.068 0.0029 -0.069 0.033 0.1 0.0029 0.036 -0.0044 -0.051 -0.044 -0.033 0.005 0.05 0.0036 -0.015 0.021 -0.033 -0.018 -0.04 0.0097 0.013 -0.026 -0.025 -0.04 -0.021 -0.03 -0.013 -0.026 -0.0045 -0.0031 -0.0015 -0.024 -0.039 -0.011 0.026 0.012 -0.0046 0.0012 0.037 -0.027 -0.0024 0.061 0.0054 0.00053 0.076 -0.0053 0.019 0.017 -0.0038 0.0045 -0.0076 0.045 0.0044 -0.0063 -0.0033 -0.0023 -0.019 -0.02 -0.0048 -0.014 -0.027 -0.016 -0.025 -0.025 0.03 -0.038 0.031 -0.04 -0.0095 -0.0051 -0.022 -0.061 -0.021 -0.044 -0.019 -0.08 -0.0073 -0.069 -0.0047 -0.014 -0.05 -0.01 -0.0096 -0.018 -0.021 -0.011 -0.03 -0.054 -0.0075 -0.026 -0.057 -0.023 -0.014 -0.029 -0.015 -0.026 -0.021 -0.019 -0.018 1.0 -0.015 -0.017 -0.023 -0.048 -0.12 -0.075 -0.051 -0.079 -0.0098 -0.019 -0.011 -0.047 -0.051 -0.025 -0.055
country_Republic of Congo 0.0016 -0.051 -0.0021 -0.02 0.052 -0.014 -0.0063 -0.013 0.014 -0.0032 0.049 -0.02 -0.0045 -0.01 -0.0057 -0.003 -0.0061 -0.0024 0.0035 -0.021 0.063 -0.0052 -0.0032 -0.00042 0.0027 0.0015 -0.0064 -0.0031 -0.0065 -0.0031 -0.00059 -0.00064 -0.013 -0.0077 0.0096 0.0035 -0.0037 -0.0016 -0.0016 0.0036 0.0011 0.0013 -0.0088 -0.00068 -0.00044 -0.0098 -0.0015 -0.005 -0.0037 -0.0011 -0.00083 -0.0021 -0.014 -0.0012 -0.0012 0.019 -0.00044 0.0039 -0.00068 -0.0011 0.019 0.00091 -0.00033 -0.004 -0.0014 -0.016 0.0099 -0.0067 -0.0075 -0.0018 -0.00096 -0.0042 -0.012 -0.004 -0.0084 -0.0036 -0.015 -0.0014 -0.013 -0.00089 -0.0027 -0.0095 -0.0019 -0.0018 -0.0034 -0.0039 -0.002 -0.0056 -0.01 -0.0014 -0.0049 -0.011 -0.0043 -0.0027 -0.0055 -0.0028 -0.0049 -0.0039 -0.0036 -0.0034 -0.015 1.0 -0.0033 -0.0044 -0.0091 -0.023 -0.014 -0.0096 -0.015 -0.0018 -0.0036 -0.0021 -0.0089 -0.0096 -0.0047 -0.01
country_Rwanda -0.029 -0.067 0.0062 0.0084 0.0091 -0.016 -0.018 -0.017 -0.00096 -0.0034 -0.0068 -0.0015 -0.0082 -0.012 -0.0076 -0.0062 0.037 -0.0017 0.0024 0.0046 -0.0018 -0.006 -0.011 -0.0089 0.04 0.0076 0.0055 -0.0036 -0.0075 -0.0036 -0.00068 -0.00074 0.012 -0.0079 -0.0058 -0.0073 -0.0043 -0.0019 -0.0018 0.018 -0.0064 -0.0022 -0.01 -0.00079 -0.0005 -0.0097 -0.0017 -0.012 -0.0057 -0.0013 -0.0045 4.7e-05 -0.015 -0.0014 -0.0014 -0.00073 -0.0005 0.0062 0.001 -0.0013 -0.0024 -0.0071 -0.0028 -0.0078 -0.01 -0.026 0.0021 0.04 -0.0087 -0.0021 -0.0011 -0.0049 -0.013 -0.0047 -0.0097 -0.0041 -0.018 -0.0016 -0.015 -0.001 -0.0031 -0.011 -0.0022 -0.0021 -0.0039 -0.0046 -0.0023 -0.0065 -0.012 -0.0017 -0.0056 -0.013 -0.005 -0.0032 -0.0064 -0.0033 -0.0056 -0.0045 -0.0042 -0.0039 -0.017 -0.0033 1.0 -0.0051 -0.011 -0.026 -0.016 -0.011 -0.017 -0.0021 -0.0042 -0.0024 -0.01 -0.011 -0.0055 -0.012
country_Senegal 0.0025 -0.02 -0.0035 0.012 -0.024 -0.017 0.013 0.047 0.0034 -0.016 0.0066 -0.0014 -0.017 -0.012 0.012 -0.0045 -0.019 -0.0077 -0.0085 0.0047 -0.019 -0.0065 0.012 0.0054 -0.012 -0.0056 -0.0053 0.043 -0.01 -0.0048 -0.00091 -0.001 0.0054 -0.012 -0.0097 -0.0053 -0.0025 -2.6e-05 -0.0024 -0.014 -0.0085 -0.003 -0.0085 0.0047 -0.00068 -0.014 -0.0023 -0.0021 0.015 -0.0018 0.0049 0.00046 0.046 -0.0019 -0.0018 0.0052 -0.00068 -0.0061 -0.0039 -0.0017 -0.011 -0.0088 -0.0046 -0.01 -0.012 0.048 -0.0099 -0.017 -0.012 -0.0028 -0.0015 -0.0066 -0.018 -0.0063 -0.013 -0.0055 -0.023 -0.0022 -0.02 -0.0014 -0.0041 -0.015 -0.003 -0.0028 -0.0053 -0.0061 -0.0031 -0.0087 -0.016 -0.0022 -0.0076 -0.017 -0.0067 -0.0042 -0.0086 -0.0044 -0.0075 -0.0061 -0.0056 -0.0052 -0.023 -0.0044 -0.0051 1.0 -0.014 -0.035 -0.022 -0.015 -0.023 -0.0029 -0.0056 -0.0033 -0.014 -0.015 -0.0073 -0.016
country_Sierra Leone -0.25 -0.34 -0.0099 0.18 0.01 -0.041 -0.047 -0.076 0.0091 0.11 -0.065 0.038 -0.04 -0.032 -0.014 -0.016 0.043 -0.017 -0.017 -0.054 -0.048 -0.014 -0.03 -0.028 -0.035 -0.013 0.22 0.0042 0.1 0.0033 -0.0019 -0.002 0.17 0.066 0.055 -0.01 -0.012 -0.0051 -0.005 -0.022 -0.0093 -0.003 -0.027 -0.0022 -0.0014 -0.03 -0.0047 -0.03 -0.012 -0.0036 -0.01 -0.0039 -0.071 -0.0039 -0.0021 -0.002 0.021 0.022 0.14 -0.0036 -0.048 0.035 0.16 0.33 -0.044 -0.096 0.08 -0.011 -0.024 -0.0057 -0.003 -0.013 -0.037 -0.013 -0.027 -0.011 -0.048 -0.0044 -0.041 -0.0028 -0.0085 -0.03 -0.0061 -0.0058 -0.011 -0.013 -0.0064 -0.018 -0.032 -0.0045 -0.016 -0.034 -0.014 -0.0087 -0.018 -0.009 -0.015 -0.012 -0.012 -0.011 -0.048 -0.0091 -0.011 -0.014 1.0 -0.072 -0.045 -0.03 -0.048 -0.0059 -0.011 -0.0067 -0.028 -0.031 -0.015 -0.033
country_Somalia -0.42 0.11 -0.015 0.0093 0.057 0.045 -0.11 -0.15 -0.0033 0.19 -0.015 0.16 0.093 0.056 -0.047 -0.037 -0.02 0.06 0.07 0.045 0.086 0.021 -0.075 -0.057 -0.059 -0.024 -0.0074 0.0011 0.073 0.03 0.0045 0.0075 -0.015 0.21 -0.021 0.02 -0.0052 -0.011 -0.0072 0.016 0.095 0.0065 0.048 -0.0028 0.0027 0.015 0.0028 -0.068 -0.033 -0.009 -0.026 -0.011 -0.14 -0.0098 -0.0056 -0.005 -0.0014 0.032 0.014 -0.0041 0.2 -0.0079 -0.01 0.026 0.12 -0.2 -0.035 -0.034 -0.059 -0.014 -0.0076 -0.034 -0.092 -0.032 -0.066 -0.028 -0.12 -0.011 -0.1 -0.0071 -0.021 -0.075 -0.015 -0.014 -0.027 -0.031 -0.016 -0.044 -0.081 -0.011 -0.039 -0.086 -0.034 -0.022 -0.044 -0.023 -0.038 -0.031 -0.029 -0.027 -0.12 -0.023 -0.026 -0.035 -0.072 1.0 -0.11 -0.076 -0.12 -0.015 -0.028 -0.017 -0.071 -0.076 -0.038 -0.082
country_South Africa 0.14 0.089 -0.015 -0.1 -0.096 -0.02 0.24 0.2 -0.0041 -0.051 0.012 -0.099 -0.073 -0.048 0.031 -0.01 -0.098 -0.018 -0.026 -0.1 -0.076 -0.022 0.14 -0.0027 -0.058 -0.02 -0.036 -0.015 -0.032 -0.015 -0.0029 -0.0032 -0.067 -0.042 -0.032 -0.022 -0.016 0.00023 -0.00027 -0.061 -0.026 -0.0074 -0.041 -0.0015 -0.0022 0.021 -0.0073 0.2 0.0036 0.0012 0.043 0.03 0.21 0.015 0.015 -0.0031 -0.0022 -0.015 -0.019 -0.0056 -0.15 -0.03 -0.018 -0.033 -0.064 0.33 -0.064 -0.083 -0.037 -0.0089 -0.0047 -0.021 -0.057 -0.02 -0.042 -0.018 -0.075 -0.0069 -0.065 -0.0044 -0.013 -0.047 -0.0096 -0.009 -0.017 -0.02 -0.0099 -0.028 -0.051 -0.0071 -0.024 -0.054 -0.021 -0.014 -0.027 -0.014 -0.024 -0.019 -0.018 -0.017 -0.075 -0.014 -0.016 -0.022 -0.045 -0.11 1.0 -0.047 -0.074 -0.0092 -0.018 -0.01 -0.044 -0.048 -0.023 -0.051
country_South Sudan 0.19 0.13 0.0037 -0.0027 -0.014 0.053 -0.048 -0.066 0.0017 -0.0054 0.0095 0.051 -0.031 0.036 -0.016 -0.017 0.032 -0.003 0.056 0.078 -0.025 0.026 -0.035 -0.028 0.057 0.017 -0.0083 0.00064 -0.007 -0.00079 -0.002 -0.0022 -0.02 -0.023 -0.014 -0.01 0.0054 -0.0053 -0.0052 -0.0042 -0.0099 0.016 0.033 -0.0023 -0.0015 0.035 7.5e-05 -0.031 -0.012 -0.0022 -0.0055 -0.00086 -0.06 -0.0041 -0.00085 0.0038 -0.0015 0.023 0.0027 -0.0021 0.058 0.018 -0.0097 -0.017 -0.013 -0.09 0.013 0.034 -0.025 -0.006 -0.0032 -0.014 -0.039 -0.013 -0.028 -0.012 -0.051 -0.0047 -0.044 -0.003 -0.0089 -0.032 -0.0064 -0.0061 -0.011 -0.013 -0.0067 -0.019 -0.034 -0.0048 -0.016 -0.036 -0.014 -0.0092 -0.018 -0.0095 -0.016 -0.013 -0.012 -0.011 -0.051 -0.0096 -0.011 -0.015 -0.03 -0.076 -0.047 1.0 -0.05 -0.0062 -0.012 -0.0071 -0.03 -0.032 -0.016 -0.035
country_Sudan 0.29 0.016 0.011 0.032 0.051 0.025 -0.065 -0.07 0.0016 -0.027 0.02 -0.023 -0.048 0.023 -0.022 -0.0017 0.1 0.0087 -0.0082 0.039 -0.054 0.00049 -0.039 0.011 0.093 0.013 -0.017 0.011 -0.0034 0.00062 -0.00095 -0.0034 -0.018 -0.025 -0.017 -0.0098 0.0048 -0.0037 -0.0034 0.08 0.0036 0.005 0.024 -0.0036 0.0034 0.012 0.0074 -0.045 -0.017 -0.0049 -0.015 -0.0081 -0.078 -0.0065 -0.0041 -0.0013 0.0063 -0.0089 -0.012 -0.0014 -0.0056 0.028 -0.013 -0.031 0.098 -0.11 -0.01 0.067 -0.039 -0.0094 -0.005 -0.022 -0.06 -0.021 -0.044 -0.019 -0.079 -0.0073 -0.068 -0.0047 -0.014 -0.049 -0.01 -0.0095 -0.018 -0.021 -0.01 -0.029 -0.053 -0.0075 -0.026 -0.057 -0.022 -0.014 -0.029 -0.015 -0.025 -0.021 -0.019 -0.018 -0.079 -0.015 -0.017 -0.023 -0.048 -0.12 -0.074 -0.05 1.0 -0.0097 -0.019 -0.011 -0.047 -0.05 -0.025 -0.054
country_Swaziland -0.051 -0.0088 -0.002 -0.014 -0.0064 -0.0086 -0.0015 0.02 -0.00054 -0.007 0.0082 -0.013 -0.01 -0.0067 3.3e-05 0.041 0.00014 -0.00027 -0.0018 -0.013 -0.0059 -0.0033 0.0062 0.045 0.018 -0.0027 -0.0047 -0.002 -0.0042 -0.002 -0.00038 -0.00042 -0.0087 -0.0055 -0.0045 -0.0041 -0.0024 -0.001 -0.001 -0.004 -0.0018 -0.0012 -0.0057 -0.00044 -0.00028 -0.0053 -0.00096 -0.007 -0.0032 -0.00074 0.0046 -0.0014 0.011 -0.0008 0.015 -0.00041 -0.00028 -0.0025 -0.0025 -0.00073 -0.02 -0.0039 -0.0024 -0.0043 -0.00037 0.023 0.011 -0.0058 -0.0048 -0.0012 -0.00062 -0.0027 -0.0075 -0.0026 -0.0054 -0.0023 -0.0098 -0.0009 -0.0084 -0.00058 -0.0017 -0.0061 -0.0012 -0.0012 -0.0022 -0.0025 -0.0013 -0.0036 -0.0066 -0.00092 -0.0032 -0.007 -0.0028 -0.0018 -0.0036 -0.0018 -0.0031 -0.0025 -0.0023 -0.0022 -0.0098 -0.0018 -0.0021 -0.0029 -0.0059 -0.015 -0.0092 -0.0062 -0.0097 1.0 -0.0023 -0.0014 -0.0058 -0.0062 -0.0031 -0.0067
country_Tanzania 0.1 -0.0075 -0.0029 -0.02 0.0094 0.013 0.026 0.0016 -0.001 -0.0098 0.0019 -0.024 -0.008 0.0014 0.0072 0.0024 0.013 -0.0065 -0.0069 -0.023 -0.0093 0.001 0.018 0.0094 0.0059 -0.003 -0.0044 -0.0039 -0.0082 -0.0039 -0.00074 -0.00081 -0.014 -0.0095 0.0038 -0.0048 -0.0046 0.046 -0.002 0.016 -0.0052 0.0077 0.0037 -0.00086 -0.00055 0.014 0.0014 0.021 -0.0023 -0.0014 0.0012 -0.0026 -0.0014 -0.0016 -0.0015 -0.00079 -0.00055 3.4e-05 -0.0048 -0.0014 -0.025 -0.0077 0.0007 -0.0084 -0.0041 0.022 -0.00043 0.012 -0.0094 -0.0023 -0.0012 -0.0053 -0.015 -0.0051 -0.011 -0.0045 -0.019 -0.0017 -0.016 -0.0011 -0.0033 -0.012 -0.0024 -0.0023 -0.0043 -0.0049 -0.0025 -0.0071 -0.013 -0.0018 -0.0061 -0.014 -0.0054 -0.0034 -0.0069 -0.0036 -0.0061 -0.0049 -0.0045 -0.0042 -0.019 -0.0036 -0.0042 -0.0056 -0.011 -0.028 -0.018 -0.012 -0.019 -0.0023 1.0 -0.0027 -0.011 -0.012 -0.006 -0.013
country_Togo -0.039 -0.0042 -0.002 -0.015 -0.015 -0.0082 0.014 0.041 -0.00061 -0.0072 0.016 -0.015 -0.0096 -0.0044 0.00015 0.016 -0.017 -0.0051 -0.004 -0.015 -0.011 -0.0038 0.019 0.037 0.00064 -0.0012 -0.0054 -0.0023 -0.0048 -0.0023 -0.00043 -0.00047 -0.0099 -0.0062 -0.004 0.00067 -0.0027 -0.0012 -0.0012 -0.012 -0.0041 -0.0014 -0.0026 -0.0005 -0.00032 -0.0072 -0.0011 0.0023 -0.0037 0.0064 0.0034 -0.0016 0.028 0.019 -0.00088 -0.00046 -0.00032 -0.0029 -0.0028 -0.00083 -0.02 -0.0045 -0.0027 -0.0049 -0.01 0.044 -0.0068 -0.011 -0.0055 -0.0013 -0.00071 -0.0031 -0.0085 -0.003 -0.0062 -0.0026 -0.011 -0.001 -0.0096 -0.00066 -0.002 -0.007 -0.0014 -0.0013 -0.0025 -0.0029 -0.0015 -0.0041 -0.0075 -0.0011 -0.0036 -0.008 -0.0032 -0.002 -0.0041 -0.0021 -0.0036 -0.0029 -0.0027 -0.0025 -0.011 -0.0021 -0.0024 -0.0033 -0.0067 -0.017 -0.01 -0.0071 -0.011 -0.0014 -0.0027 1.0 -0.0066 -0.0071 -0.0035 -0.0076
country_Tunisia -0.25 0.1 -0.0092 -0.06 -0.065 -0.041 0.054 0.18 0.0022 -0.033 -0.0096 -0.061 -0.015 -0.026 0.011 0.014 -0.092 -0.02 -0.016 -0.059 -0.0077 -0.013 0.035 0.0075 -0.032 -0.012 -0.022 -0.0096 -0.019 -0.0097 -0.0018 -0.002 -0.041 -0.026 -0.011 -0.016 -0.01 -1.3e-05 0.0015 -0.055 -0.016 -0.006 -0.022 -0.0021 -0.0014 -0.031 -0.0046 0.05 0.00076 0.0034 0.002 -0.0056 0.18 0.012 -0.002 0.0043 -0.0014 -0.011 -0.011 -0.0035 -0.076 -0.019 -0.011 -0.021 -0.018 0.19 -0.014 -0.084 -0.023 -0.0056 -0.003 -0.013 -0.036 -0.013 -0.026 -0.011 -0.047 -0.0043 -0.041 -0.0028 -0.0083 -0.03 -0.006 -0.0057 -0.011 -0.012 -0.0063 -0.018 -0.032 -0.0045 -0.015 -0.034 -0.013 -0.0085 -0.017 -0.0089 -0.015 -0.012 -0.011 -0.011 -0.047 -0.0089 -0.01 -0.014 -0.028 -0.071 -0.044 -0.03 -0.047 -0.0058 -0.011 -0.0066 1.0 -0.03 -0.015 -0.032
country_Uganda 0.0054 -0.15 -0.0034 0.12 -0.067 0.026 -0.004 -0.046 -0.0028 -0.027 0.024 0.022 -0.045 0.0052 0.016 0.0046 0.025 -0.016 -0.012 0.071 -0.043 0.056 0.01 0.0016 -0.009 -0.0058 0.057 -0.0097 -0.02 0.0051 0.0043 -0.0022 0.1 -0.025 -0.013 -0.019 -0.011 -0.00073 -0.0029 -0.049 -0.016 -0.0045 -0.011 0.0086 -0.0015 0.035 -0.0012 -0.0092 0.0034 -0.0022 0.0015 -0.0071 -0.047 0.0018 -0.0024 -0.0021 -0.0015 -0.0017 -0.0095 -0.0038 0.028 -0.017 0.0096 -0.023 -0.037 -0.035 0.017 0.031 -0.025 -0.006 -0.0032 -0.014 -0.039 -0.014 -0.028 -0.012 -0.051 -0.0047 -0.044 -0.003 -0.0089 -0.032 -0.0065 -0.0061 -0.011 -0.013 -0.0067 -0.019 -0.034 -0.0048 -0.016 -0.036 -0.014 -0.0092 -0.019 -0.0095 -0.016 -0.013 -0.012 -0.011 -0.051 -0.0096 -0.011 -0.015 -0.031 -0.076 -0.048 -0.032 -0.05 -0.0062 -0.012 -0.0071 -0.03 1.0 -0.016 -0.035
country_Zambia 0.092 -0.005 -0.0049 -0.033 -0.012 -0.016 0.07 0.036 -0.0014 -0.014 -0.0043 -0.033 -0.017 -0.017 0.021 -0.0025 -0.0074 -0.0092 -0.0077 -0.034 -0.023 -0.0085 0.03 0.0036 0.0032 -0.0061 -0.012 -0.0051 -0.011 -0.0051 -0.00097 -0.0011 -0.021 -0.013 -0.0052 0.0027 -0.0061 -0.00035 -0.00024 -0.0012 -0.0091 -0.0032 -0.014 -0.0011 -0.00072 -0.005 -0.0025 0.063 0.014 0.0046 0.025 0.0017 0.036 -0.0021 -0.002 -0.001 -0.00072 0.003 -0.0064 -0.0019 -0.048 -0.01 -0.005 -0.011 -0.022 0.076 -0.0044 -0.0059 -0.012 -0.003 -0.0016 -0.007 -0.019 -0.0067 -0.014 -0.0059 -0.025 -0.0023 -0.022 -0.0015 -0.0044 -0.016 -0.0032 -0.003 -0.0056 -0.0065 -0.0033 -0.0093 -0.017 -0.0024 -0.0081 -0.018 -0.0071 -0.0045 -0.0091 -0.0047 -0.008 -0.0065 -0.006 -0.0056 -0.025 -0.0047 -0.0055 -0.0073 -0.015 -0.038 -0.023 -0.016 -0.025 -0.0031 -0.006 -0.0035 -0.015 -0.016 1.0 -0.017
country_Zimbabwe -0.092 -0.11 -0.011 -0.076 0.17 -0.049 -0.02 -0.028 -0.003 -0.039 -0.049 -0.073 -0.045 -0.037 0.0039 0.001 0.17 -0.024 -0.017 -0.074 -0.057 -0.018 -0.016 0.012 0.11 -0.015 -0.026 -0.011 -0.024 -0.011 -0.0021 -0.0023 -0.049 -0.031 0.042 -0.0012 -0.013 -0.0025 0.0043 0.21 -0.02 -0.0069 -0.032 -0.0025 -0.0016 -0.032 -0.0042 -0.019 0.014 -0.0041 -0.0098 -0.0068 -0.035 -0.0031 -0.0014 -0.0023 -0.0016 -0.013 -0.014 -0.0041 -0.11 -0.022 0.048 -0.024 -0.048 -0.037 -0.0048 0.19 -0.027 -0.0065 -0.0035 -0.015 -0.042 -0.015 -0.03 -0.013 -0.055 -0.005 -0.047 -0.0032 -0.0096 -0.034 -0.007 -0.0066 -0.012 -0.014 -0.0073 -0.02 -0.037 -0.0052 -0.018 -0.039 -0.016 -0.0099 -0.02 -0.01 -0.018 -0.014 -0.013 -0.012 -0.055 -0.01 -0.012 -0.016 -0.033 -0.082 -0.051 -0.035 -0.054 -0.0067 -0.013 -0.0076 -0.032 -0.035 -0.017 1.0
In [87]:
median_fatal=np.median(features.fatalities)
median_fatal
Out[87]:
0.0
In [88]:
mean_fatal=np.mean(features.fatalities)
mean_fatal
Out[88]:
4.420257164913635
In [89]:
features_small['deadly']=[1 if i > median_fatal else 0 for i in features_small.fatalities]
In [90]:
features_small['deadly'].value_counts()
Out[90]:
0    22639
1    10523
Name: deadly, dtype: int64
In [91]:
target=features_small.deadly
In [92]:
#  Baseline using the smaller dataset
1-features_small['deadly'].mean()
Out[92]:
0.6826789699053133
In [93]:
#Baseline calculated with the full dataset:
baseline=(56660/(56660+26244))
print('Baseline Accuracy: ', baseline)

Baseline Accuracy:  0.6834410884878896

This is the baseline calculated assuming the whole dataset. We get the exact same value so the random sample is representative of dataset.

baseline=(113321/(113321+52487)) print('Baseline Accuracy: ', baseline)

I'm not using the fatalities for obvious reasons so i will drop them at this stage.

In [94]:
features_small.drop(['fatalities', 'deadly'], axis=1, inplace=True)
In [95]:
features_small.shape
Out[95]:
(33162, 115)
In [96]:
features_small.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 33162 entries, 148345 to 98715
Columns: 115 entries, neighbours to country_Zimbabwe
dtypes: int64(2), uint8(113)
memory usage: 4.3 MB
In [97]:
from sklearn.preprocessing import StandardScaler
scaler = StandardScaler()
categorical_ss= scaler.fit_transform(features_small)
In [98]:
from sklearn.model_selection import train_test_split
f_train, f_test, t_train, t_test = train_test_split(features_small, target, random_state=1)

Feature Selection

At this stage, running the models are taking an inordinate amount of time (time i no longer have to spend) so in order to carry out the ML part of the project im going to randomly resamople the data.

https://machinelearningmastery.com/feature-selection-to-improve-accuracy-and-decrease-training-time/

In [99]:
import sys
from yellowbrick.features import Rank1D

Method 1:

In [100]:
from yellowbrick.features import Rank1D
In [101]:
from sklearn.ensemble import GradientBoostingClassifier

from yellowbrick.features.importances import FeatureImportances

# Create a new matplotlib figure
fig = plt.figure(figsize=(18,16))
ax = fig.add_subplot()

viz = FeatureImportances(GradientBoostingClassifier(), ax=ax)
viz.fit(f_train, t_train)
viz.poof()

Okay, a little side-note. I tried different techniques to identify the features to drop and they generally agreed. the one that intuitevely made the most sense was this first method because it identifies the weakest features as small countries with few events and the interactions which are a poor feature so I'm happy to drop them. The section below is the removal of the weakest features.

In [102]:
feat_2 = pd.DataFrame({'feature':viz.features_,'importance':viz.feature_importances_})
feat_2.set_index('feature',inplace=True)
In [103]:
feat_2.sort_values('importance', ascending=False)
Out[103]:
importance
feature
year 100.000000
event_type_Riots/Protests 91.031089
event_type_Strategic development 51.740244
neighbours 42.834958
event_type_Battle-No change of territory 38.523412
country_Sierra Leone 38.038967
country_Zimbabwe 35.745965
country_South Sudan 31.185549
country_Liberia 28.916007
event_type_Non-violent transfer of territory 27.885365
interaction_12 27.363399
interaction_30 25.747466
inter2_7 24.502753
inter1_4 24.314548
country_Nigeria 21.300365
event_type_Violence against civilians 20.709891
country_Ethiopia 19.658244
inter1_6 19.197178
interaction_57 18.257646
country_Angola 17.605785
event_type_Remote violence 16.730562
inter1_2 15.984975
country_Egypt 15.756518
interaction_27 15.221322
interaction_60 13.948171
country_Sudan 13.305024
country_Somalia 10.236073
interaction_15 8.319370
interaction_78 8.191345
interaction_37 6.374815
... ...
interaction_18 0.000000
country_Burkina Faso 0.000000
country_Burundi 0.000000
interaction_26 0.000000
interaction_25 0.000000
country_Chad 0.000000
interaction_56 0.000000
inter1_5 0.000000
interaction_16 0.000000
country_Swaziland 0.000000
interaction_66 0.000000
country_Libya 0.000000
country_Gabon 0.000000
interaction_14 0.000000
country_Ghana 0.000000
country_Guinea 0.000000
inter2_8 0.000000
country_Guinea-Bissau 0.000000
inter2_6 0.000000
country_Lesotho 0.000000
country_Madagascar 0.000000
country_Equatorial Guinea 0.000000
inter2_2 0.000000
country_Mali 0.000000
country_Mauritania 0.000000
inter1_7 0.000000
country_Mozambique 0.000000
country_Namibia 0.000000
country_Senegal 0.000000
interaction_68 0.000000

115 rows × 1 columns

In [104]:
to_exclude_1 = feat_2.loc[feat_2['importance']<1]
In [105]:
to_exclude_1.shape
Out[105]:
(64, 1)
In [106]:
drop_list=list(to_exclude_1.T)
In [107]:
drop_list
Out[107]:
['interaction_68',
 'interaction_22',
 'country_Chad',
 'interaction_25',
 'interaction_26',
 'country_Burundi',
 'country_Burkina Faso',
 'country_Botswana',
 'interaction_33',
 'country_Benin',
 'country_Djibouti',
 'interaction_35',
 'country_Tanzania',
 'country_Togo',
 'interaction_40',
 'interaction_45',
 'interaction_46',
 'event_type_Battle-Non-state actor overtakes territory',
 'interaction_77',
 'interaction_70',
 'interaction_55',
 'interaction_36',
 'interaction_56',
 'interaction_18',
 'interaction_16',
 'country_Swaziland',
 'country_Senegal',
 'country_Namibia',
 'country_Mozambique',
 'inter1_5',
 'inter1_7',
 'country_Mauritania',
 'country_Mali',
 'inter2_2',
 'country_Equatorial Guinea',
 'country_Madagascar',
 'country_Lesotho',
 'inter2_6',
 'country_Guinea-Bissau',
 'inter2_8',
 'country_Guinea',
 'country_Ghana',
 'interaction_14',
 'country_Gabon',
 'country_Libya',
 'interaction_66',
 'interaction_24',
 'interaction_44',
 'country_Morocco',
 'interaction_11',
 'interaction_80',
 'interaction_23',
 'interaction_28',
 'country_Uganda',
 'country_Eritrea',
 'country_Gambia',
 'interaction_88',
 'country_Malawi',
 'country_Cameroon',
 'interaction_50',
 'country_Kenya',
 'interaction_48',
 'country_Tunisia',
 'interaction_38']

Method 2:

In [108]:
forest = ExtraTreesClassifier(n_estimators=250,
                              random_state=0)
forest.fit(f_train, t_train)
importances = forest.feature_importances_
std = np.std([tree.feature_importances_ for tree in forest.estimators_],
             axis=0)
indices = np.argsort(importances)[::-1]
In [109]:
print("Feature ranking:")

for f in range(f_train.shape[1]):
    print("%d. feature %d (%f)" % (f + 1, indices[f], importances[indices[f]]))

# Plot the feature importances of the forest
plt.figure()
plt.title("Feature importances")
plt.bar(range(f_train.shape[1]), importances[indices],
       color="r", yerr=std[indices], align="center")
plt.xticks(range(f_train.shape[1]), indices)
plt.xlim([-1, f_train.shape[1]])
plt.show()

Feature ranking:
1. feature 1 (0.266079)
2. feature 64 (0.064878)
3. feature 59 (0.052832)
4. feature 65 (0.047424)
5. feature 6 (0.043355)
6. feature 51 (0.041104)
7. feature 66 (0.037879)
8. feature 114 (0.029572)
9. feature 0 (0.025797)
10. feature 15 (0.024428)
11. feature 103 (0.022329)
12. feature 4 (0.017551)
13. feature 63 (0.013121)
14. feature 5 (0.012233)
15. feature 104 (0.012144)
16. feature 99 (0.012056)
17. feature 107 (0.011405)
18. feature 18 (0.011328)
19. feature 44 (0.011006)
20. feature 106 (0.010639)
21. feature 80 (0.010620)
22. feature 75 (0.009383)
23. feature 62 (0.008011)
24. feature 33 (0.007822)
25. feature 2 (0.007670)
26. feature 77 (0.006761)
27. feature 89 (0.006761)
28. feature 46 (0.006666)
29. feature 3 (0.006320)
30. feature 10 (0.005874)
31. feature 105 (0.005819)
32. feature 19 (0.005641)
33. feature 12 (0.005640)
34. feature 31 (0.005559)
35. feature 71 (0.005387)
36. feature 73 (0.005376)
37. feature 38 (0.005312)
38. feature 9 (0.004960)
39. feature 67 (0.004895)
40. feature 90 (0.004881)
41. feature 25 (0.004817)
42. feature 11 (0.004812)
43. feature 86 (0.004393)
44. feature 93 (0.004261)
45. feature 60 (0.004093)
46. feature 112 (0.003936)
47. feature 87 (0.003798)
48. feature 23 (0.003232)
49. feature 101 (0.003090)
50. feature 8 (0.003049)
51. feature 111 (0.002828)
52. feature 41 (0.002764)
53. feature 21 (0.002441)
54. feature 72 (0.002434)
55. feature 84 (0.002345)
56. feature 113 (0.002261)
57. feature 22 (0.002206)
58. feature 98 (0.002184)
59. feature 100 (0.002149)
60. feature 96 (0.002046)
61. feature 102 (0.002003)
62. feature 32 (0.001820)
63. feature 74 (0.001788)
64. feature 16 (0.001676)
65. feature 92 (0.001658)
66. feature 27 (0.001497)
67. feature 39 (0.001349)
68. feature 109 (0.001266)
69. feature 20 (0.001253)
70. feature 49 (0.001174)
71. feature 70 (0.001169)
72. feature 56 (0.001163)
73. feature 17 (0.001162)
74. feature 83 (0.001136)
75. feature 13 (0.001112)
76. feature 35 (0.001110)
77. feature 34 (0.001103)
78. feature 91 (0.001103)
79. feature 97 (0.001101)
80. feature 24 (0.001020)
81. feature 61 (0.001020)
82. feature 95 (0.000939)
83. feature 57 (0.000925)
84. feature 28 (0.000868)
85. feature 14 (0.000859)
86. feature 82 (0.000804)
87. feature 79 (0.000797)
88. feature 45 (0.000746)
89. feature 76 (0.000667)
90. feature 68 (0.000647)
91. feature 85 (0.000604)
92. feature 40 (0.000565)
93. feature 47 (0.000531)
94. feature 110 (0.000509)
95. feature 26 (0.000499)
96. feature 94 (0.000431)
97. feature 108 (0.000344)
98. feature 81 (0.000272)
99. feature 88 (0.000270)
100. feature 50 (0.000219)
101. feature 36 (0.000214)
102. feature 78 (0.000197)
103. feature 58 (0.000158)
104. feature 37 (0.000139)
105. feature 69 (0.000077)
106. feature 29 (0.000070)
107. feature 30 (0.000055)
108. feature 53 (0.000053)
109. feature 42 (0.000052)
110. feature 43 (0.000050)
111. feature 48 (0.000046)
112. feature 7 (0.000028)
113. feature 52 (0.000011)
114. feature 55 (0.000007)
115. feature 54 (0.000004)
In [110]:
features_ET = pd.DataFrame({'variable':features_small.columns,
              'feature_importance':importances})
In [111]:
result= features_ET.sort_values('feature_importance', ascending=False)
result
Out[111]:
feature_importance variable
1 0.266079 year
64 0.064878 event_type_Riots/Protests
59 0.052832 event_type_Battle-No change of territory
65 0.047424 event_type_Strategic development
6 0.043355 inter1_6
51 0.041104 interaction_60
66 0.037879 event_type_Violence against civilians
114 0.029572 country_Zimbabwe
0 0.025797 neighbours
15 0.024428 inter2_7
103 0.022329 country_Sierra Leone
4 0.017551 inter1_4
63 0.013121 event_type_Remote violence
5 0.012233 inter1_5
104 0.012144 country_Somalia
99 0.012056 country_Nigeria
107 0.011405 country_Sudan
18 0.011328 interaction_12
44 0.011006 interaction_47
106 0.010639 country_South Sudan
80 0.010620 country_Ethiopia
75 0.009383 country_Democratic Republic of Congo
62 0.008011 event_type_Non-violent transfer of territory
33 0.007822 interaction_30
2 0.007670 inter1_2
77 0.006761 country_Egypt
89 0.006761 country_Liberia
46 0.006666 interaction_50
3 0.006320 inter1_3
10 0.005874 inter2_2
... ... ...
82 0.000804 country_Gambia
79 0.000797 country_Eritrea
45 0.000746 interaction_48
76 0.000667 country_Djibouti
68 0.000647 country_Benin
85 0.000604 country_Guinea-Bissau
40 0.000565 interaction_40
47 0.000531 interaction_55
110 0.000509 country_Togo
26 0.000499 interaction_22
94 0.000431 country_Mauritania
108 0.000344 country_Swaziland
81 0.000272 country_Gabon
88 0.000270 country_Lesotho
50 0.000219 interaction_58
36 0.000214 interaction_35
78 0.000197 country_Equatorial Guinea
58 0.000158 interaction_88
37 0.000139 interaction_36
69 0.000077 country_Botswana
29 0.000070 interaction_25
30 0.000055 interaction_26
53 0.000053 interaction_68
42 0.000052 interaction_45
43 0.000050 interaction_46
48 0.000046 interaction_56
7 0.000028 inter1_7
52 0.000011 interaction_66
55 0.000007 interaction_77
54 0.000004 interaction_70

115 rows × 2 columns

Method 3:

I tried RFE but its taking an inordinate amount of time to run so i'm including the code (which should work) but i have commented it out to avoid running it.

In [112]:
# from sklearn.svm import LinearSVC
# from sklearn.feature_selection import RFE
# svm = LinearSVC()

# rfe = RFE(svm, 10)
# rfe = rfe.fit(f_train, t_train)

# print(rfe.support_)
# print(rfe.ranking_)
In [113]:
# names=features_small.columns
# print('Features sorted by their rank:') #"Features sorted by their rank:"

# print(sorted(zip(map(lambda x: round(x,4), rfe.ranking_), names)))

Okay so now i drop the columns i have selected using drop_list

In [114]:
data=features_small.drop(drop_list, axis=1)
data.shape
Out[114]:
(33162, 51)
In [115]:
target.shape
Out[115]:
(33162,)

Modelling

Do a split of data on the new dataset.

In [116]:
X_train, X_test, y_train, y_test = train_test_split(data, target, random_state=1)
In [117]:
from sklearn.linear_model import LogisticRegression
logreg = LogisticRegression()
logreg.fit(X_train, y_train)
print('Accuracy of Logistic regression classifier on training set: {:.2f}'
     .format(logreg.score(X_train, y_train)))
print('Accuracy of Logistic regression classifier on test set: {:.2f}'
     .format(logreg.score(X_test, y_test)))

Accuracy of Logistic regression classifier on training set: 0.75
Accuracy of Logistic regression classifier on test set: 0.75
In [118]:
from sklearn.tree import DecisionTreeClassifier
clf = DecisionTreeClassifier().fit(X_train, y_train)
print('Accuracy of Decision Tree classifier on training set: {:.2f}'
     .format(clf.score(X_train, y_train)))
print('Accuracy of Decision Tree classifier on test set: {:.2f}'
     .format(clf.score(X_test, y_test)))

Accuracy of Decision Tree classifier on training set: 0.84
Accuracy of Decision Tree classifier on test set: 0.77
In [119]:
from sklearn.neighbors import KNeighborsClassifier
knn = KNeighborsClassifier()
knn.fit(X_train, y_train)
print('Accuracy of K-NN classifier on training set: {:.2f}'
     .format(knn.score(X_train, y_train)))
print('Accuracy of K-NN classifier on test set: {:.2f}'
     .format(knn.score(X_test, y_test)))

Accuracy of K-NN classifier on training set: 0.80
Accuracy of K-NN classifier on test set: 0.76
In [120]:
from sklearn.discriminant_analysis import LinearDiscriminantAnalysis
lda = LinearDiscriminantAnalysis()
lda.fit(X_train, y_train)
print('Accuracy of LDA classifier on training set: {:.2f}'.format(lda.score(X_train, y_train)))
print('Accuracy of LDA classifier on test set: {:.2f}'.format(lda.score(X_test, y_test)))

Accuracy of LDA classifier on training set: 0.75
Accuracy of LDA classifier on test set: 0.75
In [121]:
from sklearn.naive_bayes import GaussianNB
gnb = GaussianNB()
gnb.fit(X_train, y_train)
print('Accuracy of GNB classifier on training set: {:.2f}'
     .format(gnb.score(X_train, y_train)))
print('Accuracy of GNB classifier on test set: {:.2f}'
     .format(gnb.score(X_test, y_test)))

Accuracy of GNB classifier on training set: 0.69
Accuracy of GNB classifier on test set: 0.68
In [124]:
from sklearn.svm import SVC
svm = SVC(probability=True)
svm.fit(X_train, y_train)
print('Accuracy of SVM classifier on training set: {:.2f}'
     .format(svm.score(X_train, y_train)))
print('Accuracy of SVM classifier on test set: {:.2f}'
     .format(svm.score(X_test, y_test)))

Accuracy of SVM classifier on training set: 0.77
Accuracy of SVM classifier on test set: 0.76
In [125]:
from sklearn.neural_network import MLPClassifier
from sklearn import metrics

nnet = MLPClassifier(hidden_layer_sizes=5,solver='lbfgs',activation='identity',max_iter=1000,random_state=1)
nnet.fit(X_train,y_train)
metrics.mean_squared_error(y_test,nnet.predict(X_test))
print('Accuracy of NN classifier on training set: {:.2f}'
     .format(nnet.score(X_train, y_train)))
print('Accuracy of NN classifier on test set: {:.2f}'
     .format(nnet.score(X_test, y_test)))

Accuracy of NN classifier on training set: 0.68
Accuracy of NN classifier on test set: 0.68
In [126]:
from sklearn.ensemble import RandomForestClassifier
rf = RandomForestClassifier()
rf.fit(X_train, y_train)
print('Accuracy of RF classifier on training set: {:.2f}'
     .format(rf.score(X_train, y_train)))
print('Accuracy of RF classifier on test set: {:.2f}'
     .format(rf.score(X_test, y_test)))

Accuracy of RF classifier on training set: 0.84
Accuracy of RF classifier on test set: 0.77

GridSearchCV for top 3 models

First up is KNN - My best performer :)

In [127]:
from sklearn.model_selection import GridSearchCV
In [128]:
param_grid = [{'weights': ["uniform", "distance"], 'n_neighbors': [10, 15, 25]}]

# knn_clf = KNeighborsClassifier()
grid_search_knn = GridSearchCV(knn, param_grid, cv=5, verbose=3, n_jobs=-1)
grid_search_knn.fit(X_train, y_train)

Fitting 5 folds for each of 6 candidates, totalling 30 fits
[CV] n_neighbors=10, weights=uniform .................................
[CV] n_neighbors=10, weights=uniform .................................
[CV] n_neighbors=10, weights=uniform .................................
[CV] n_neighbors=10, weights=uniform .................................
[CV]  n_neighbors=10, weights=uniform, score=0.7700502512562815, total=   1.4s
[CV] n_neighbors=10, weights=uniform .................................
[CV]  n_neighbors=10, weights=uniform, score=0.7573869346733668, total=   1.4s
[CV]  n_neighbors=10, weights=uniform, score=0.7710092480900683, total=   1.4s
[CV] n_neighbors=10, weights=distance ................................
[CV] n_neighbors=10, weights=distance ................................
[CV]  n_neighbors=10, weights=uniform, score=0.7533172496984318, total=   1.3s
[CV] n_neighbors=10, weights=distance ................................
[CV]  n_neighbors=10, weights=distance, score=0.772462311557789, total=   1.7s
[CV]  n_neighbors=10, weights=uniform, score=0.7657349688316911, total=   1.7s
[CV]  n_neighbors=10, weights=distance, score=0.7595979899497487, total=   1.7s
[CV] n_neighbors=10, weights=distance ................................
[CV] n_neighbors=10, weights=distance ................................
[CV] n_neighbors=15, weights=uniform .................................
[CV]  n_neighbors=10, weights=distance, score=0.7726176115802171, total=   1.7s
[CV] n_neighbors=15, weights=uniform .................................
[CV]  n_neighbors=10, weights=distance, score=0.7643273677860446, total=   1.7s
[CV] n_neighbors=15, weights=uniform .................................
[CV]  n_neighbors=10, weights=distance, score=0.7503015681544029, total=   1.7s
[CV] n_neighbors=15, weights=uniform .................................
[CV]  n_neighbors=15, weights=uniform, score=0.7682412060301508, total=   1.8s
[CV] n_neighbors=15, weights=uniform .................................
[CV]  n_neighbors=15, weights=uniform, score=0.7553768844221106, total=   1.8s
[CV] n_neighbors=15, weights=distance ................................
[CV]  n_neighbors=15, weights=uniform, score=0.7724165661439485, total=   1.6s
[CV] n_neighbors=15, weights=distance ................................
[CV]  n_neighbors=15, weights=uniform, score=0.7595496582227583, total=   1.6s
[CV] n_neighbors=15, weights=distance ................................
[CV]  n_neighbors=15, weights=uniform, score=0.7647295395133722, total=   1.6s
[CV] n_neighbors=15, weights=distance ................................
[CV]  n_neighbors=15, weights=distance, score=0.7710552763819095, total=   1.5s
[CV] n_neighbors=15, weights=distance ................................
[CV]  n_neighbors=15, weights=distance, score=0.761608040201005, total=   1.5s
[CV] n_neighbors=25, weights=uniform .................................
[CV]  n_neighbors=15, weights=distance, score=0.7792521109770808, total=   1.4s
[CV] n_neighbors=25, weights=uniform .................................
[CV]  n_neighbors=15, weights=distance, score=0.7595496582227583, total=   1.3s
[CV] n_neighbors=25, weights=uniform .................................
[CV]  n_neighbors=15, weights=distance, score=0.7671425698773376, total=   1.3s
[CV] n_neighbors=25, weights=uniform .................................
[CV]  n_neighbors=25, weights=uniform, score=0.7666331658291458, total=   1.5s
[CV] n_neighbors=25, weights=uniform .................................
[CV]  n_neighbors=25, weights=uniform, score=0.7597989949748744, total=   1.6s
[CV] n_neighbors=25, weights=distance ................................
[CV]  n_neighbors=25, weights=uniform, score=0.7627663852030558, total=   1.5s
[CV] n_neighbors=25, weights=distance ................................
[CV]  n_neighbors=25, weights=uniform, score=0.7551266586248492, total=   1.6s
[CV] n_neighbors=25, weights=distance ................................
[CV]  n_neighbors=25, weights=uniform, score=0.7703599436959582, total=   1.8s
[CV] n_neighbors=25, weights=distance ................................
[CV]  n_neighbors=25, weights=distance, score=0.7764824120603016, total=   1.7s
[CV] n_neighbors=25, weights=distance ................................
[CV]  n_neighbors=25, weights=distance, score=0.7666331658291458, total=   1.6s
[CV]  n_neighbors=25, weights=distance, score=0.7764374748693205, total=   1.6s
[CV]  n_neighbors=25, weights=distance, score=0.7591475673502212, total=   2.1s
[CV]  n_neighbors=25, weights=distance, score=0.769555600241303, total=   2.1s

[Parallel(n_jobs=-1)]: Done  30 out of  30 | elapsed:   54.6s finished
Out[128]:
GridSearchCV(cv=5, error_score='raise',
       estimator=KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',
           metric_params=None, n_jobs=1, n_neighbors=5, p=2,
           weights='uniform'),
       fit_params=None, iid=True, n_jobs=-1,
       param_grid=[{'weights': ['uniform', 'distance'], 'n_neighbors': [10, 15, 25]}],
       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',
       scoring=None, verbose=3)
In [129]:
from sklearn.model_selection import cross_val_score
np.mean(cross_val_score(grid_search_knn.best_estimator_, X_test, y_test))
Out[129]:
0.7564823151495914
In [130]:
final_classifier = grid_search_knn.best_estimator_.fit(X_test, y_test)
final_classifier
Out[130]:
KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',
           metric_params=None, n_jobs=1, n_neighbors=25, p=2,
           weights='distance')
In [131]:
print('Accuracy of K-NN classifier on test set: {:.2f}'
     .format(final_classifier.score(X_test, y_test)))

Accuracy of K-NN classifier on test set: 0.87
In [132]:
from sklearn.metrics import classification_report
from sklearn.metrics import confusion_matrix
pred = final_classifier.predict(X_test)
print(confusion_matrix(y_test, pred))
print(classification_report(y_test, pred))

[[5289  371]
 [ 682 1949]]
             precision    recall  f1-score   support

          0       0.89      0.93      0.91      5660
          1       0.84      0.74      0.79      2631

avg / total       0.87      0.87      0.87      8291

conda install -c conda-forge scikit-plot

In [133]:
import scikitplot as skplt
skplt.metrics.plot_confusion_matrix(y_test, pred, normalize=True)
plt.show()
In [134]:
from sklearn import metrics
In [135]:
y_probas = final_classifier.predict_proba(X_test)
skplt.metrics.plot_precision_recall(y_test, y_probas)
plt.show()
In [136]:
y_probas = final_classifier.predict_proba(X_test)
skplt.metrics.plot_roc(y_test, y_probas)
plt.show()
In [137]:
y_probas = final_classifier.predict_proba(X_test)
skplt.metrics.plot_ks_statistic(y_test, y_probas)
plt.show()
In [138]:
y_probas = final_classifier.predict_proba(X_test)
skplt.metrics.plot_cumulative_gain(y_test, y_probas)
plt.show()

http://mlwiki.org/index.php/Cumulative_Gain_Chart

Next up is SVC

In [ ]:
param_grid_svm = [{'kernel':('linear', 'rbf'), 'C': [1, 10, ]}]
grid_search_svc=GridSearchCV(svm, param_grid_svm, cv=5, verbose=3)
grid_search_svc.fit(X_train, y_train)

Fitting 5 folds for each of 4 candidates, totalling 20 fits
[CV] C=1, kernel=linear ..............................................
[CV] ..... C=1, kernel=linear, score=0.7495477386934674, total= 2.7min
[CV] C=1, kernel=linear ..............................................

[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:  2.9min remaining:    0.0s

[CV] ...... C=1, kernel=linear, score=0.737286432160804, total= 2.9min
[CV] C=1, kernel=linear ..............................................

[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:  5.9min remaining:    0.0s
In [ ]:
np.mean(cross_val_score(grid_search_svc.best_estimator_, X_test, y_test))
In [ ]:
final_classifier_svc = grid_search_svc.best_estimator_.fit(X_test, y_test)
final_classifier_svc
In [ ]:
print('Accuracy of SVC classifier on test set: {:.2f}'
     .format(final_classifier_svc.score(X_test, y_test)))
In [ ]:
pred_svc = final_classifier_svc.predict(X_test)
print(confusion_matrix(y_test, pred_svc))
print(classification_report(y_test, pred_svc))
In [ ]:
skplt.metrics.plot_confusion_matrix(y_test, pred, normalize=True)
plt.show()

y_probas_svc = final_classifier_svc.predict_proba(X_test) skplt.metrics.plot_precision_recall(y_test, y_probas_svc) plt.show() AttributeError: predict_proba is not available when probability=False https://github.com/scikit-learn/scikit-learn/issues/4791

In [ ]:
# y_probas_svc = final_classifier_svc.predict_proba(X_test)
# skplt.metrics.plot_precision_recall(y_test, y_probas_svc)
# plt.show()
In [ ]:
# skplt.metrics.plot_ks_statistic(y_test, y_probas)
# plt.show()
In [ ]:
# y_probas = final_classifier_svc.predict_proba(X_test)
# skplt.metrics.plot_cumulative_gain(y_test, y_probas)
# plt.show()

And now Random Forest

In [139]:
param_grid = { 
    'n_estimators': [100],
    'max_features': ['auto', 'sqrt', 'log2'],
    'max_depth': [10, 20, 30, 40]
}

CV_rfc = GridSearchCV(rf, param_grid=param_grid, cv= 5)
CV_rfc.fit(X_train, y_train)
Out[139]:
GridSearchCV(cv=5, error_score='raise',
       estimator=RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',
            max_depth=None, max_features='auto', max_leaf_nodes=None,
            min_impurity_decrease=0.0, min_impurity_split=None,
            min_samples_leaf=1, min_samples_split=2,
            min_weight_fraction_leaf=0.0, n_estimators=10, n_jobs=1,
            oob_score=False, random_state=None, verbose=0,
            warm_start=False),
       fit_params=None, iid=True, n_jobs=1,
       param_grid={'n_estimators': [100], 'max_features': ['auto', 'sqrt', 'log2'], 'max_depth': [10, 20, 30, 40]},
       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',
       scoring=None, verbose=0)
In [140]:
# print CV_rfc.best_params_
In [141]:
np.mean(cross_val_score(CV_rfc.best_estimator_, X_test, y_test))
Out[141]:
0.7662527308086545
In [142]:
final_classifier_rf = CV_rfc.best_estimator_.fit(X_test, y_test)
final_classifier_rf
Out[142]:
RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',
            max_depth=20, max_features='sqrt', max_leaf_nodes=None,
            min_impurity_decrease=0.0, min_impurity_split=None,
            min_samples_leaf=1, min_samples_split=2,
            min_weight_fraction_leaf=0.0, n_estimators=100, n_jobs=1,
            oob_score=False, random_state=None, verbose=0,
            warm_start=False)
In [143]:
print('Accuracy of RF classifier on test set: {:.2f}'
     .format(final_classifier_rf.score(X_test, y_test)))

Accuracy of RF classifier on test set: 0.87
In [146]:
pred_rf = final_classifier_rf.predict(X_test)
print(confusion_matrix(y_test, pred_rf))
print(classification_report(y_test, pred_rf))

[[5152  508]
 [ 602 2029]]
             precision    recall  f1-score   support

          0       0.90      0.91      0.90      5660
          1       0.80      0.77      0.79      2631

avg / total       0.87      0.87      0.87      8291
In [147]:
skplt.metrics.plot_confusion_matrix(y_test, pred_rf, normalize=True)
plt.show()
In [148]:
y_probas_rf = final_classifier_rf.predict_proba(X_test)
skplt.metrics.plot_precision_recall(y_test, y_probas_rf)
plt.show()
In [149]:
y_probas_rf = final_classifier_rf.predict_proba(X_test)
skplt.metrics.plot_roc(y_test, y_probas_rf)
plt.show()
In [150]:
y_probas_rf = final_classifier_rf.predict_proba(X_test)
skplt.metrics.plot_ks_statistic(y_test, y_probas_rf)
plt.show()
In [151]:
y_probas_rf = final_classifier_rf.predict_proba(X_test)
skplt.metrics.plot_cumulative_gain(y_test, y_probas_rf)
plt.show()

Last one is the Decison tree which i will bag first

In [152]:
from sklearn.ensemble import BaggingClassifier
In [153]:
model = BaggingClassifier(base_estimator=clf, n_estimators=10)
params = {'max_samples': np.linspace(0.8, 1.0, 3),
          'max_features': range(int(3/4.*data.shape[1]), data.shape[1]+1)}

grid = GridSearchCV(model, param_grid=params, cv=5)
grid.fit(X_train, y_train)
Out[153]:
GridSearchCV(cv=5, error_score='raise',
       estimator=BaggingClassifier(base_estimator=DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=None,
            max_features=None, max_leaf_nodes=None,
            min_impurity_decrease=0.0, min_impurity_split=None,
            min_samples_leaf=1, min_samples_split=2,
            ...n_estimators=10, n_jobs=1, oob_score=False,
         random_state=None, verbose=0, warm_start=False),
       fit_params=None, iid=True, n_jobs=1,
       param_grid={'max_samples': array([0.8, 0.9, 1. ]), 'max_features': range(38, 52)},
       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',
       scoring=None, verbose=0)
In [154]:
np.mean(cross_val_score(grid.best_estimator_, X_test, y_test))
Out[154]:
0.755639044579682
In [155]:
final_classifier_dt = grid.best_estimator_.fit(X_test, y_test)
final_classifier_dt
Out[155]:
BaggingClassifier(base_estimator=DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=None,
            max_features=None, max_leaf_nodes=None,
            min_impurity_decrease=0.0, min_impurity_split=None,
            min_samples_leaf=1, min_samples_split=2,
            min_weight_fraction_leaf=0.0, presort=False, random_state=None,
            splitter='best'),
         bootstrap=True, bootstrap_features=False, max_features=39,
         max_samples=1.0, n_estimators=10, n_jobs=1, oob_score=False,
         random_state=None, verbose=0, warm_start=False)
In [156]:
print('Accuracy of DT classifier on test set: {:.2f}'
     .format(final_classifier_dt.score(X_test, y_test)))

Accuracy of DT classifier on test set: 0.86
In [157]:
pred_dt = final_classifier_dt.predict(X_test)
print(confusion_matrix(y_test, pred_dt))
print(classification_report(y_test, pred_dt))

[[5102  558]
 [ 643 1988]]
             precision    recall  f1-score   support

          0       0.89      0.90      0.89      5660
          1       0.78      0.76      0.77      2631

avg / total       0.85      0.86      0.85      8291
In [158]:
skplt.metrics.plot_confusion_matrix(y_test, pred_dt, normalize=True)
plt.show()
In [159]:
y_probas_dt = final_classifier_dt.predict_proba(X_test)
skplt.metrics.plot_precision_recall(y_test, y_probas_dt)
plt.show()
In [160]:
y_probas_dt = final_classifier_dt.predict_proba(X_test)
skplt.metrics.plot_roc(y_test, y_probas_dt)
plt.show()
In [161]:
y_probas_dt = final_classifier_dt.predict_proba(X_test)
skplt.metrics.plot_ks_statistic(y_test, y_probas_dt)
plt.show()
In [162]:
y_probas_dt = final_classifier_dt.predict_proba(X_test)
skplt.metrics.plot_cumulative_gain(y_test, y_probas_dt)
plt.show()
In [163]:
feature_importances = np.mean([
    tree.feature_importances_ for tree in final_classifier_dt.estimators_
], axis=0)
In [164]:
features = list(X_test.columns)
In [165]:
fi=pd.DataFrame(list(zip(features,feature_importances)))
In [174]:
fi.columns=['features', 'feature_importances']
fi.head()
Out[174]:
features feature_importances
0 neighbours 0.027360
1 year 0.014315
2 inter1_2 0.015651
3 inter1_3 0.005486
4 inter1_4 0.047534
In [175]:
fi.sort_values(by=['feature_importances'])

# >>> df.sort_values(by=['col1'])
Out[175]:
features feature_importances
3 inter1_3 0.005486
35 country_Democratic Republic of Congo 0.008134
27 event_type_Headquarters or base established 0.008990
37 country_Ethiopia 0.009051
33 country_Angola 0.009547
9 inter2_4 0.010077
20 interaction_37 0.011951
18 interaction_30 0.013184
25 interaction_78 0.013743
1 year 0.014315
26 event_type_Battle-No change of territory 0.015011
31 event_type_Strategic development 0.015052
5 inter1_6 0.015349
2 inter1_2 0.015651
10 inter2_5 0.016578
24 interaction_60 0.018360
32 event_type_Violence against civilians 0.020905
38 country_Ivory Coast 0.021145
36 country_Egypt 0.021579
19 interaction_34 0.023437
14 interaction_15 0.024113
21 interaction_47 0.024206
7 inter2_1 0.024937
6 inter1_8 0.026757
30 event_type_Riots/Protests 0.027305
0 neighbours 0.027360
28 event_type_Non-violent transfer of territory 0.029536
17 interaction_27 0.030085
23 interaction_58 0.034639
16 interaction_20 0.034959
8 inter2_3 0.037774
29 event_type_Remote violence 0.038821
12 interaction_12 0.044232
13 interaction_13 0.044553
4 inter1_4 0.047534
22 interaction_57 0.051443
15 interaction_17 0.054233
34 country_Central African Republic 0.054293
11 inter2_7 0.055675

I'm commenting this out also because i generated the tree and its massive and takes far too long to run

In [ ]:
# from sklearn.externals.six import StringIO  
# from IPython.display import Image  
# from sklearn.tree import export_graphviz
# import pydotplus
In [ ]:
# dot_data = StringIO()
# export_graphviz(clf, out_file=dot_data,  
#                 filled=True, rounded=True,
#                 special_characters=True)
# graph = pydotplus.graph_from_dot_data(dot_data.getvalue())  
# Image(graph.create_png())
In [ ]:
# graph.write_png('./dtr1.png')

Okay so this final step will allow me to compare the different models together

In [176]:
probas_list=[y_probas_dt, y_probas, y_probas_rf]
class_names=['Decision Trees', 'KNN', 'Random Forest']
In [177]:
skplt.metrics.plot_calibration_curve(y_test, probas_list, class_names)
plt.show()

When performing classification you often want not only to predict the class label, but also obtain a probability of the respective label. This probability gives you some kind of confidence on the prediction. Some models can give you poor estimates of the class probabilities and some even do not not support probability prediction. The calibration module allows you to better calibrate the probabilities of a given model, or to add support for probability prediction.

Well calibrated classifiers are probabilistic classifiers for which the output of the predict_proba method can be directly interpreted as a confidence level. For instance, a well calibrated (binary) classifier should classify the samples such that among the samples to which it gave a predict_proba value close to 0.8, approximately 80% actually belong to the positive class.

http://scikit-learn.org/stable/auto_examples/calibration/plot_calibration_curve.html

http://scikit-plot.readthedocs.io/en/stable/Quickstart.html

In [178]:
model_names=['Logistic Regression', 'Decision Tree Classifier', 'KNeighbours Classifier', 'Linear Discriminant Analysis', 'GaussianNB', 'Support Vector Machines', 'Neural Networks', 'Random Forest']
In [179]:
training_scores=[0.75, 0.84, 0.80, 0.75, 0.69, 0.77, 0.68, 0.84]
In [180]:
test_scores=[0.75, 0.77, 0.76, 0.75, 0.68, 0.76, 0.68, 0.78]
In [181]:
GS_scores=[0, 0.86, 0.87, 0, 0, 0.8, 0, 0.87]
In [189]:
df=pd.DataFrame()
In [190]:
df['Training Scores']=training_scores
df['Test Scores']=test_scores
df['GridSearch']=GS_scores
df['Classifier']=model_names
In [191]:
df.head(2)
Out[191]:
Training Scores Test Scores GridSearch Classifier
0 0.75 0.75 0.00 Logistic Regression
1 0.84 0.77 0.86 Decision Tree Classifier
In [192]:
df.set_index('Classifier', inplace=True)
df['Classifier_Names']=model_names
In [193]:
df
Out[193]:
Training Scores Test Scores GridSearch Classifier_Names
Classifier
Logistic Regression 0.75 0.75 0.00 Logistic Regression
Decision Tree Classifier 0.84 0.77 0.86 Decision Tree Classifier
KNeighbours Classifier 0.80 0.76 0.87 KNeighbours Classifier
Linear Discriminant Analysis 0.75 0.75 0.00 Linear Discriminant Analysis
GaussianNB 0.69 0.68 0.00 GaussianNB
Support Vector Machines 0.77 0.76 0.80 Support Vector Machines
Neural Networks 0.68 0.68 0.00 Neural Networks
Random Forest 0.84 0.78 0.87 Random Forest
In [194]:
df.plot(x='Classifier_Names', y=['Training Scores', 'Test Scores', 'GridSearch'], kind='bar')

/Users/alexandra/anaconda3/lib/python3.6/site-packages/pandas/plotting/_core.py:1716: UserWarning: Pandas doesn't allow columns to be created via a new attribute name - see https://pandas.pydata.org/pandas-docs/stable/indexing.html#attribute-access
  series.name = label
Out[194]:
<matplotlib.axes._subplots.AxesSubplot at 0x1a38bae2b0>
In [ ]:
fig, ax = plt.subplots(figsize=(10,8))

sns.countplot(conflicts['neighbours'], label='Count', ax=ax)
plt.xticks(rotation=90)
plt.show()

overfitting. This seems to be an issue as my top 3 performers seem to have overfit. The grid search seems to have done a good job at correcting that.

Time Series Analysis

In [198]:
tsa=conflicts[['year', 'event_date', 'fatalities']]
tsa.head(3)
Out[198]:
year event_date fatalities
0 2001 18/04/2001 1
1 2001 19/04/2001 0
2 2001 20/04/2001 0
In [199]:
tsa['deadly']= [1 if i > 0 else 0 for i in conflicts.fatalities]

/Users/alexandra/anaconda3/lib/python3.6/site-packages/ipykernel_launcher.py:1: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy
  """Entry point for launching an IPython kernel.
In [200]:
tsa.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 165808 entries, 0 to 165807
Data columns (total 4 columns):
year          165808 non-null int64
event_date    165808 non-null object
fatalities    165808 non-null int64
deadly        165808 non-null int64
dtypes: int64(3), object(1)
memory usage: 11.3+ MB
In [201]:
tsa1=tsa.sample(frac=0.2, random_state=1)
In [202]:
tsa1['event_date'] = pd.to_datetime(tsa.event_date)
In [203]:
tsa1.dtypes
Out[203]:
year                   int64
event_date    datetime64[ns]
fatalities             int64
deadly                 int64
dtype: object
In [204]:
tsa1['month']=tsa1.event_date.dt.month
tsa1.month.head()
Out[204]:
148345     3
45752     11
107284    10
80089      5
165648     1
Name: month, dtype: int64
In [210]:
tsa1 = tsa1.rename(columns={'year': 'y'})
In [212]:
tsa1.head(3)
Out[212]:
y event_date fatalities deadly month ds
148345 2008 2008-03-26 1 1 3 26
45752 2002 2002-11-12 9 1 11 12
107284 2007 2007-10-08 4 1 10 8
In [ ]:
from fbprophet import Prophet
In [307]:
tsa_2['month'] = tsa_1.resample('M').count().month
In [308]:
tsa_2.plot(logy=True)
Out[308]:
<matplotlib.axes._subplots.AxesSubplot at 0x1a2bdac9e8>
In [309]:
tsa_2.head()
Out[309]:
fatalities deadly month
event_date
1997-01-31 335 18 78
1997-02-28 227 14 28
1997-03-31 163 18 53
1997-04-30 767 9 38
1997-05-31 1260 12 60

Prophet is a procedure for forecasting time series data. It is based on an additive model where non-linear trends are fit with yearly and weekly seasonality, plus holidays. Prophet is designed for analyzing time series with daily observations that display patterns on different time scales. https://github.com/facebook/prophet/tree/master/python

In [213]:
from fbprophet import Prophet
In [230]:
# m = Prophet()
# m.fit(tsa1)  
# future = m.make_future_dataframe(periods=365)
# m.predict(future)

Unfortunately this returned MemoryError and i dont have time to try anything else!!

A tiny taster of NLP (but not really :))

In [215]:
from wordcloud import WordCloud, STOPWORDS
In [216]:
text=conflicts.notes
text[1]
Out[216]:
'Riots were reported in numerous villages in Kabylie, resulting in dozens wounded in clashes between protesters and police and significant material damage.'
In [217]:
text.isnull().sum()
Out[217]:
10227
In [218]:
type(text)
Out[218]:
pandas.core.series.Series
In [219]:
text=text.dropna()
In [220]:
text.isnull().sum()
Out[220]:
0
In [221]:
len(text)
Out[221]:
155581
In [222]:
stopwords= set(STOPWORDS)
In [223]:
stopwords.update('MDC')
In [224]:
tokens = []
for word in text:
    words=word.lower().split()
    tokens.extend(words)
In [225]:
tokens
Out[225]:
['a',
 'berber',
 'student',
 'was',
 'shot',
 'while',
 'in',
 'police',
 'custody',
 'at',
 'a',
 'police',
 'station',
 'in',
 'beni',
 'douala.',
 'he',
 'later',
 'died',
 'on',
 'apr.21.',
 'riots',
 'were',
 'reported',
 'in',
 'numerous',
 'villages',
 'in',
 'kabylie,',
 'resulting',
 'in',
 'dozens',
 'wounded',
 'in',
 'clashes',
 'between',
 'protesters',
 'and',
 'police',
 'and',
 'significant',
 'material',
 'damage.',
 'students',
 'protested',
 'in',
 'the',
 'amizour',
 'area.',
 'at',
 'least',
 '3',
 'were',
 'later',
 'arrested',
 'for',
 'allegedly',
 'insulting',
 'gendarmes.',
 'rioters',
 'threw',
 'molotov',
 'cocktails,',
 'rocks',
 'and',
 'burning',
 'tires',
 'at',
 'gendarmerie',
 'stations',
 'in',
 'beni',
 'douala,',
 'el-kseur',
 'and',
 'amizour.',
 'rioters',
 'threw',
 'molotov',
 'cocktails,',
 'rocks',
 'and',
 'burning',
 'tires',
 'at',
 'gendarmerie',
 'stations',
 'in',
 'beni',
 'douala,',
 'el-kseur',
 'and',
 'amizour.',
 'rioters',
 'threw',
 'molotov',
 'cocktails,',
 'rocks',
 'and',
 'burning',
 'tires',
 'at',
 'gendarmerie',
 'stations',
 'in',
 'beni',
 'douala,',
 'el-kseur',
 'and',
 'amizour.',
 'protesters',
 'attacked',
 'a',
 'gendarmerie',
 'detachment',
 'with',
 'rocks',
 'and',
 'set',
 'fire',
 'to',
 'two',
 'gendarmerie',
 'vehicle',
 'as',
 'well',
 'as',
 'the',
 'daira,',
 'registry',
 'office',
 'and',
 'court',
 'buildings.',
 'five',
 'people',
 'were',
 'killed',
 'in',
 '3',
 'days',
 'of',
 'rioting',
 'across',
 'the',
 'kabyle',
 'region',
 'after',
 'a',
 'student',
 'died',
 'in',
 'police',
 'custody',
 'in',
 'beni',
 'douala.',
 'protesters',
 'threw',
 'stones',
 'and',
 'set',
 'fire',
 'to',
 'public',
 'buildings.',
 'five',
 'people',
 'were',
 'killed',
 'in',
 '3',
 'days',
 'of',
 'rioting',
 'across',
 'the',
 'kabyle',
 'region',
 'after',
 'a',
 'student',
 'died',
 'in',
 'police',
 'custody',
 'in',
 'beni',
 'douala.',
 'protesters',
 'threw',
 'stones',
 'and',
 'set',
 'fire',
 'to',
 'public',
 'buildings.',
 'clashes',
 'between',
 'police',
 'and',
 'protesters',
 'werereported',
 'in',
 'sidi',
 'aich,',
 'el',
 'kseur,',
 'tazmalt,',
 'barbacha,',
 'seddouk',
 'and',
 'timezrit.',
 'various',
 'arson',
 'attacks',
 'were',
 'reported.',
 'clashes',
 'between',
 'police',
 'and',
 'protesters',
 'werereported',
 'in',
 'sidi',
 'aich,',
 'el',
 'kseur,',
 'tazmalt,',
 'barbacha,',
 'seddouk',
 'and',
 'timezrit.',
 'various',
 'arson',
 'attacks',
 'were',
 'reported.',
 'clashes',
 'between',
 'police',
 'and',
 'protesters',
 'werereported',
 'in',
 'sidi',
 'aich,',
 'el',
 'kseur,',
 'tazmalt,',
 'barbacha,',
 'seddouk',
 'and',
 'timezrit.',
 'various',
 'arson',
 'attacks',
 'were',
 'reported.',
 'clashes',
 'between',
 'police',
 'and',
 'protesters',
 'werereported',
 'in',
 'sidi',
 'aich,',
 'el',
 'kseur,',
 'tazmalt,',
 'barbacha,',
 'seddouk',
 'and',
 'timezrit.',
 'various',
 'arson',
 'attacks',
 'were',
 'reported.',
 'clashes',
 'between',
 'police',
 'and',
 'protesters',
 'werereported',
 'in',
 'sidi',
 'aich,',
 'el',
 'kseur,',
 'tazmalt,',
 'barbacha,',
 'seddouk',
 'and',
 'timezrit.',
 'various',
 'arson',
 'attacks',
 'were',
 'reported.',
 'clashes',
 'between',
 'police',
 'and',
 'protesters',
 'werereported',
 'in',
 'sidi',
 'aich,',
 'el',
 'kseur,',
 'tazmalt,',
 'barbacha,',
 'seddouk',
 'and',
 'timezrit.',
 'various',
 'arson',
 'attacks',
 'were',
 'reported.',
 'five',
 'people',
 'were',
 'killed',
 'in',
 '3',
 'days',
 'of',
 'rioting',
 'across',
 'the',
 'kabyle',
 'region',
 'after',
 'a',
 'student',
 'died',
 'in',
 'police',
 'custody',
 'in',
 'beni',
 'douala.',
 'protesters',
 'threw',
 'stones',
 'and',
 'set',
 'fire',
 'to',
 'public',
 'buildings.',
 '15',
 'people',
 'were',
 'killed',
 'between',
 'apr.25-26',
 'in',
 'a',
 'continuation',
 'of',
 'rioting',
 'from',
 'earlier',
 'this',
 'week',
 'across',
 'the',
 'kabyle',
 'region',
 'following',
 'the',
 'death',
 'of',
 'a',
 'student',
 'in',
 'police',
 'custody',
 'in',
 'beni',
 'douala.',
 'protesters',
 'threw',
 'stones',
 'and',
 'set',
 'fire',
 'to',
 'public',
 'buildings.',
 'the',
 'offices',
 'of',
 'the',
 'rcd',
 'and',
 'ffs',
 'were',
 'set',
 'on',
 'fire',
 'in',
 'the',
 'tizi',
 'rached',
 'area',
 'during',
 'ongoing',
 'unrest',
 'in',
 'the',
 'kabyle',
 'region.',
 '15',
 'people',
 'were',
 'killed',
 'between',
 'apr.25-26',
 'in',
 'a',
 'continuation',
 'of',
 'rioting',
 'from',
 'earlier',
 'this',
 'week',
 'across',
 'the',
 'kabyle',
 'region',
 'following',
 'the',
 'death',
 'of',
 'a',
 'student',
 'in',
 'police',
 'custody',
 'in',
 'beni',
 'douala.',
 'protesters',
 'threw',
 'stones',
 'and',
 'set',
 'fire',
 'to',
 'public',
 'buildings.',
 '9',
 'people',
 'were',
 'killed,',
 'including',
 'a',
 'police',
 'officer,',
 'and',
 '19',
 'injured',
 'in',
 'clashes',
 'between',
 'berber',
 'protesters',
 'and',
 'police',
 'in',
 'tizi',
 'ouzou',
 'and',
 'villages',
 'around',
 'it,',
 'as',
 'well',
 'as',
 'bejaia,',
 'following',
 'the',
 'death',
 'of',
 'a',
 'student',
 'in',
 'police',
 'custody.',
 '29',
 'people',
 'have',
 'been',
 'killed',
 'since',
 'monday',
 'in',
 'rioting',
 'across',
 'the',
 'region.',
 'betwee',
 '40-60',
 'security',
 'personnel',
 'were',
 'killed',
 'in',
 'a',
 'clash',
 'with',
 'an',
 'unidentified',
 'group',
 'to',
 'the',
 'south',
 'of',
 'tebessa.',
 '9',
 'people',
 'were',
 'killed,',
 'including',
 'a',
 'police',
 'officer,',
 'and',
 '19',
 'injured',
 'in',
 'clashes',
 'between',
 'berber',
 'protesters',
 'and',
 'police',
 'in',
 'tizi',
 'ouzou',
 'and',
 'villages',
 'around',
 'it,',
 'as',
 'well',
 'as',
 'bejaia,',
 'following',
 'the',
 'death',
 'of',
 'a',
 'student',
 'in',
 'police',
 'custody.',
 '29',
 'people',
 'have',
 'been',
 'killed',
 'since',
 'monday',
 'in',
 'rioting',
 'across',
 'the',
 'region.',
 '9',
 'people',
 'were',
 'killed,',
 'including',
 'a',
 'police',
 'officer,',
 'and',
 '19',
 'injured',
 'in',
 'clashes',
 'between',
 'berber',
 'protesters',
 'and',
 'police',
 'in',
 'tizi',
 'ouzou',
 'and',
 'villages',
 'around',
 'it,',
 'as',
 'well',
 'as',
 'bejaia,',
 'following',
 'the',
 'death',
 'of',
 'a',
 'student',
 'in',
 'police',
 'custody.',
 '29',
 'people',
 'have',
 'been',
 'killed',
 'since',
 'monday',
 'in',
 'rioting',
 'across',
 'the',
 'region.',
 'berber',
 'protesters',
 'and',
 'police',
 'fought',
 'running',
 'battles',
 'in',
 'tizi',
 'ouzou',
 'and',
 'bejaia.',
 'at',
 'least',
 '13',
 'were',
 'killed',
 'on',
 'apr.28.',
 'at',
 'least',
 '42',
 'people',
 'reported',
 'killed',
 'and',
 'over',
 '400',
 'injured',
 'in',
 'total,',
 'mostly',
 'police,',
 'since',
 'apr.23',
 'according',
 'to',
 'official',
 'figures.',
 'berber',
 'protesters',
 'and',
 'police',
 'fought',
 'running',
 'battles',
 'in',
 'tizi',
 'ouzou',
 'and',
 'bejaia.',
 'at',
 'least',
 '13',
 'were',
 'killed',
 'on',
 'apr.28.',
 'at',
 'least',
 '42',
 'people',
 'reported',
 'killed',
 'and',
 'over',
 '400',
 'injured',
 'in',
 'total,',
 'mostly',
 'police,',
 'since',
 'apr.23',
 'according',
 'to',
 'official',
 'figures.',
 'separately',
 'from',
 'a',
 'sit-in',
 'staged',
 'the',
 'same',
 'day,',
 'between',
 '10-20,000',
 'people',
 'marched',
 'in',
 'algiers',
 'under',
 'the',
 'banner',
 'of',
 'the',
 'front',
 'for',
 'socialist',
 'forces',
 'against',
 'recent',
 'violence',
 'in',
 'the',
 'kabyle',
 'region.',
 'there',
 'were',
 'no',
 'reports',
 'of',
 'violence.',
 'about',
 '5,000',
 'people',
 'staged',
 'a',
 'peaceful',
 'anti-government',
 'protest',
 'organized',
 'by',
 'the',
 'rcd',
 'in',
 'algiers.',
 'they',
 'engaged',
 'in',
 'a',
 'sit-in',
 'outside',
 'the',
 "pm's",
 'office',
 'while',
 'riot',
 'police',
 'and',
 'helicopters',
 'were',
 'deployed',
 'but',
 'did',
 'not',
 'interact',
 'with',
 'the',
 'protesters.',
 'separately',
 'from',
 'the',
 'sit-in,',
 'between',
 '10-20,000',
 'people',
 'marched',
 'in',
 'algiers',
 'under',
 'the',
 'banner',
 'of',
 'the',
 'front',
 'for',
 'socialist',
 'forces',
 'against',
 'recent',
 'violence',
 'in',
 'the',
 'kabyle',
 'region.',
 'there',
 'were',
 'no',
 'reports',
 'of',
 'violence.',
 'students',
 'demonstrated',
 'in',
 'oran',
 'despite',
 'significant',
 'police',
 'presence.',
 'as',
 'many',
 'as',
 '10,000',
 'students',
 'protested',
 'at',
 'the',
 'government',
 'palace',
 'in',
 'algiers.',
 'no',
 'incidents',
 'reported.',
 'large',
 'protests',
 'were',
 'reported',
 'in',
 'bejaia',
 'in',
 'the',
 'context',
 'of',
 'ongoing',
 'unrest',
 'in',
 "algeria's",
 'berber',
 'regions.',
 'about',
 '200',
 'lawyers',
 'marched',
 'to',
 'the',
 'bejaia',
 'city',
 'hall',
 'to',
 'show',
 'support',
 'for',
 'those',
 'killed',
 'recently',
 'in',
 'kabyle.',
 'sporadic',
 'protests',
 'and',
 'clashes',
 'were',
 'reported,',
 'centring',
 'mainly',
 'in',
 'the',
 'bejaia',
 'region,',
 'including',
 'youths',
 'barricading',
 'roads',
 'with',
 'trees',
 'and',
 'tires',
 'set',
 'on',
 'fire.',
 'dozens',
 'of',
 'people',
 'were',
 'said',
 'to',
 'have',
 'been',
 'injured,',
 'although',
 'officially',
 'there',
 'were',
 'no',
 'injuries.',
 'the',
 'protests',
 'took',
 'place',
 'in',
 'el',
 'keur,',
 'oued',
 'ghir,',
 'bakaro,',
 'tichy,',
 'soul',
 'el',
 'thenine,',
 'bordj',
 'mira,',
 'kerrata,',
 'and',
 'aokas.',
 'sporadic',
 'protests',
 'and',
 'clashes',
 'were',
 'reported,',
 'centring',
 'mainly',
 'in',
 'the',
 'bejaia',
 'region,',
 'including',
 'youths',
 'barricading',
 'roads',
 'with',
 'trees',
 'and',
 'tires',
 'set',
 'on',
 'fire.',
 'dozens',
 'of',
 'people',
 'were',
 'said',
 'to',
 'have',
 'been',
 'injured,',
 'although',
 'officially',
 'there',
 'were',
 'no',
 'injuries.',
 'the',
 'protests',
 'took',
 'place',
 'in',
 'el',
 'keur,',
 'oued',
 'ghir,',
 'bakaro,',
 'tichy,',
 'soul',
 'el',
 'thenine,',
 'bordj',
 'mira,',
 'kerrata,',
 'and',
 'aokas.',
 'sporadic',
 'protests',
 'and',
 'clashes',
 'were',
 'reported,',
 'centring',
 'mainly',
 'in',
 'the',
 'bejaia',
 'region,',
 'including',
 'youths',
 'barricading',
 'roads',
 'with',
 'trees',
 'and',
 ...]
In [226]:
tokens_clean=[word for word in tokens if word not in stopwords]
In [247]:
wordcloud2=WordCloud(background_color='white',width=5800,
                          height=3000,max_words=100, relative_scaling=0.8, stopwords=stopwords).generate(str(text))
In [248]:
print(wordcloud2)

<wordcloud.wordcloud.WordCloud object at 0x1a56e7b898>
In [249]:
fig = plt.figure(figsize=(15,8))
plt.imshow(wordcloud, interpolation='bilinear')
plt.axis('off')
plt.show()
fig.savefig("word1.png", dpi=1200)
In [237]:
from collections import Counter
counts=Counter(tokens_clean)
In [238]:
counts.most_common(20)
Out[238]:
[('killed', 26468),
 ('police', 22761),
 ('forces', 20893),
 ('two', 17686),
 ('protest', 14108),
 ('people', 13358),
 ('attack', 12247),
 ('government', 11674),
 ('soldiers', 10586),
 ('attacked', 10273),
 ('al', 9839),
 ('one', 9707),
 ('group', 9595),
 ('military', 9288),
 ('armed', 9128),
 ('shabaab', 8701),
 ('rebels', 8595),
 ('near', 8195),
 ('security', 8059),
 ('three', 7925)]
In [239]:
from textblob import TextBlob, Word
import re
In [240]:
string= text.to_string()
In [241]:
blob=TextBlob(string)
In [242]:
blob.sentiment
Out[242]:
Sentiment(polarity=-0.07327669779923575, subjectivity=0.2522611370287405)